As of September 14, 2026, not one hurricane had formed in the Atlantic. Only five relatively weak and short-lived tropical storms had spun up all season, and the National Hurricane Center saw none likely for at least another week, even though this is normally the peak. On September 12, 2026 the season set the record for the longest a season has ever begun without a storm reaching 74 mph, at least in the satellite era, according to Nick Novella of the Weather Prediction Center.

Overall activity, measured in a way that combines storm count, strength and duration, is at a post-1950 record low. Colorado State University’s Phil Klotzbach puts Atlantic storm activity at 7% of normal for the date.

The cause is a very large El Niño, and the mechanism is wind shear rather than cool water. In the corridor between Africa and the Caribbean where the strongest storms usually form, shear is running roughly 45 mph above what a tropical cyclone can withstand, a level Kristen Corbosiero of the University at Albany described as unheard of. This is happening despite record-hot oceans; the world’s seas have been at record warmth for 100 days. El Niño has simply overwhelmed the fuel.

That distinction matters for what comes next. Kerry Emanuel of MIT, quoted on September 14, 2026, explained that Atlantic waters lag the atmosphere by several months in feeling El Niño, so next year, as the El Niño winds down and the atmosphere cools toward normal, the warm water will persist — which increases the odds that the 2027 season is extra busy. The one Atlantic area where shear has stayed low is the Gulf of Mexico, where storms tend to form late in the season.

Reinsurance pricing has been moving the other way for two years. The Guy Carpenter US property catastrophe rate-on-line index fell 14% at the April 2026 renewals, and June 2026 renewals produced risk-adjusted declines of 15% to 20%. Moody’s notes property catastrophe pricing has fallen more than 20% in the 18 months since 2024 and expects further softening at January 1, 2027; Fitch expects terms and conditions to loosen as well; S&P expects softening to continue through 2027. Analysts cited by the reinsurance trade press put another 10% to 15% risk-adjusted decline on the table for January 1 if major losses stay limited. Munich Re and Hannover Re have both publicly said they expect prices and terms to be broadly upheld at that renewal.

The capital doing the pricing is at record levels. Catastrophe bond issuance reached $17.6 billion in the first half of 2026, the largest half-year on record, and the outstanding market closed June 2026 at a record $65.6 billion, up from $61.3 billion at the end of 2025.

Share prices verified at the September 15, 2026 close: RenaissanceRe $328.70, 2.7% below its one-year high; Everest Group $379.55, 4.8% below; Arch Capital $97.04, 8.9% below; Munich Re €512.80; Hannover Re €255.40.

Opportunity

The obvious reading is that a loss-free year is good for reinsurers and that cheaper reinsurance is good for the insurers who buy it. Both are true for 2026. What the reading leaves out is that catastrophe reinsurance is priced once a year, in advance, and the January 1, 2027 renewal is where 2027’s hurricane risk gets sold. The evidence being used to set that price is a season whose quietness has a single, identifiable and explicitly temporary cause.

El Niño suppresses Atlantic hurricanes by tearing them apart, not by removing the heat that powers them. The heat is at a record. When the shear goes, the fuel is still there — which is precisely what Emanuel described, and the pattern that has historically followed strong El Niño events as they decay toward neutral or La Niña conditions.

Hypothesis: the market is extrapolating a benign loss year into a benign risk year, and the reinsurance cycle’s own machinery — record capital, record catastrophe bond issuance, three consecutive years of rate declines — will push January 2027 pricing to a cycle low at the exact moment the physical driver of 2026’s quiet reverses. If that is right, whoever writes 2027 US wind at those prices is taking the other side of a forecast they are not being paid for.

This cuts in the opposite direction for the buyers. A third consecutive year of falling catastrophe reinsurance cost is a direct reduction in the largest expense line of every Florida and Gulf Coast primary insurer, and that arrives whether or not a storm does.

How it could play out

El Niño peaks over the 2026-27 northern winter and decays through spring. Reinsurers renew on January 1, 2027 at another double-digit risk-adjusted decline with looser terms, on the strength of two light loss years. Catastrophe bond spreads tighten further as capital keeps arriving. By June 2027 the shear is gone, the water is still warm, and seasonal forecasters publish an above-normal outlook into a market that has just sold the risk at the cheapest price of the cycle. Either nothing lands, in which case 2028 pricing falls again and the cycle extends, or something does, into books written at the bottom with the least protective terms. The asymmetry belongs to whoever is short that risk while it is cheap and long it afterwards.

Questions worth asking

• Does the historical record actually support “strong El Niño year, then busy Atlantic year,” or is that a story built on a handful of cases? Counting the post-1950 transitions honestly is what decides this idea, because everything else follows from it.
• Who benefits from cheap reinsurance rather than being hurt by it? Florida and Gulf primary insurers buy catastrophe cover for June 1 each year, and a third year of falling prices is a measurable margin input rather than a risk.
• If reinsurers can read the same forecasts, why would they write it? Is there a structural reason — capital that must be deployed, fee income earned on assets under management, market share defence — that makes writing at the wrong price rational for an individual firm?
• Where does the risk actually end up? Catastrophe bond and insurance-linked securities investors have taken a growing share of peak US wind, so a loss might land on fund investors rather than on listed reinsurer balance sheets.
• The Gulf of Mexico is the one Atlantic region where shear has stayed low, and Gulf storms form late. Does the 2026 season still carry a tail, and would a late Gulf landfall reverse the January pricing conversation outright?
• El Niño is doing the opposite in the Pacific, where the eastern basin has run at nearly double normal activity. Is exposure to Hawaii, Mexico’s Pacific coast or Central America mispriced in the other direction?

Where to look

• RenaissanceRe, Everest Group, Arch Capital — Bermuda property catastrophe writers whose January renewal book is the direct expression, and whose shares sit within single digits of one-year highs after a loss-free season
• Munich Re and Hannover Re — both have said publicly they expect to hold pricing at January 2027, which makes them the test of whether underwriting discipline or capital supply wins
• Lancashire, Hiscox, Beazley — London market names with concentrated catastrophe exposure on smaller balance sheets
• Universal Insurance, HCI Group, Heritage Insurance — Florida primaries on the buying side, for whom falling reinsurance cost is an input rather than a hazard
• the catastrophe bond and insurance-linked securities market, where record issuance and record outstanding capital are the mechanism actually setting the price
• Florida’s Citizens Property Insurance and the state’s hurricane catastrophe fund, as the place residual risk collects when private capital prices it away

Thesis check

The facts underneath are hard and unusually well documented: a satellite-era record in the Atlantic with a named physical cause, and a reinsurance pricing cycle that three rating agencies have independently forecast to keep softening into January 2027, while reinsurer shares trade near one-year highs. The weakness is that “El Niño fades, so next season is busier” is a probabilistic statement about a whole ocean basin rather than a forecast of landfall, and a reinsurer can lose money for years waiting for a storm or make money for years while none arrives — so this is a claim about price versus risk, not a dated event, and it has no natural stop.

Timing

January 1, 2027 reinsurance renewals; the 2027 Atlantic season opens June 1, 2027

Sources

Insurance Journal / Associated Press, Sep 14 2026 · Artemis.bm, reinsurance softening to continue in 2027 (S&P), 2026 · Artemis.bm, catastrophe bond records broken in H1 2026 · Insurance Business, Moody's on further renewal softening, 2026 · Artemis.bm, Munich Re on January 2027 renewals, 2026

Deputy Secretary of War Steve Feinberg signed a memorandum titled “Fostering One Strong Industrial Base” on September 15, 2026, with an implementation appendix setting deadlines for the department’s senior leaders. The department describes the regime it is dismantling in unusually plain terms: a wall built “from cost accounting rules, duplicative audits, and compliance obligations that attach to entire companies rather than the work procured,” built to control contractor costs, which instead “limited who competes and raised prices.”

The directed actions are specific. The department will propose to the Cost Accounting Standards Board that exemption become the default, with coverage confined to cost-based development contracts awarded without adequate competition. It will immediately use the higher cost-accounting thresholds enacted in the fiscal year 2026 National Defense Authorization Act. Any acquisition strategy that would bring a new business unit under full coverage now requires senior-level approval. Commercial product and service determinations must be made within 15 business days. Business-system criteria become commercial-aligned, with certification by an independent public accounting firm accepted in place of separate government review. Audits become risk-based. Other transactions and advance market commitments get expanded use.

Two further elements matter more than the accounting mechanics. The memorandum implements Executive Order 14402, which makes fixed-price contracts the default and cost-reimbursement the exception. And it directs a profit policy “under which negotiated margins reflect value delivered, risk carried, and private capital invested — not merely cost incurred.”

It also carries forward the department’s August 18, 2026 direction on supplier transparency, so that contracting officers can see through a prime contractor to its suppliers when needed, and connect those suppliers directly to Office of Strategic Capital loans, Industrial Base Analysis and Sustainment funds, and the department’s business operators. No new organisation or compliance framework is created, and implementing guidance may not add requirements beyond the memorandum and applicable law.

Share prices verified at the September 15, 2026 close: Northrop Grumman $531.25, 30.8% below its one-year high; L3Harris $249.66, 34.0% below; Lockheed Martin $533.46, 21.2% below; RTX $195.50, 13.3% below; General Dynamics $358.60, 9.4% below.

Opportunity

Read as procurement reform, this is one more memorandum in a genre that rarely produces much, and that is roughly how the defense trade press has treated it. What that reading misses is which barrier is being removed, and who it was protecting.

Cost Accounting Standards coverage, government audit of business systems, and government-unique compliance obligations attach to a company, not to a contract. A commercial manufacturer that wins a single defense award has historically had to rebuild its accounting function around rules nobody else in the economy uses, and keep it that way. That entry cost is the reason most of American manufacturing does not bid — and it is therefore the moat around the firms that have already paid it.

“Attach rules to contracts, not companies” removes the entry cost. “Rely first on the independent audits companies already pay for” removes the duplicate. Paying margin for “risk carried and private capital invested” rewrites the arithmetic that has for decades made a sole-source cost-plus program a low-risk annuity and a fixed-price program a hazard to be avoided.

Hypothesis: if this is implemented as written, the effect is not that defense budgets grow but that the share of them captured by incumbent primes falls, and the character of prime earnings changes — more fixed-price execution risk, more performers carried to production, less protected sole-source sustainment, and contracting officers looking straight past the prime at its suppliers and financing them directly. The sub-tier suppliers whose margins are currently set by primes sit on the other side of every one of those clauses. None of this appears in anyone’s model, because the mechanism is an accounting standard rather than a program.

How it could play out

The department proposes the Cost Accounting Standards exemption change and begins using the higher statutory thresholds immediately. Commercial determinations inside 15 business days make it practical for a company with no defense accounting apparatus to sell a product. More performers get carried to production, so program awards stop being winner-take-all. Fixed-price default moves execution risk onto contractors, widening the gap between those who can absorb it and those who cannot. Suppliers that previously reached the department only through a prime are seen directly and financed directly through Office of Strategic Capital loans. The prime contractor’s three historic functions — compliance apparatus, systems integration, and supply chain financier — are each unbundled by a different clause, and margin follows risk and capital to whoever is actually carrying them.

Questions worth asking

• Does the Cost Accounting Standards Board actually adopt the proposal? It is a statutory body outside the Department of War, the change has been resisted for decades, and without it most of this remains aspiration. This single question decides the idea.
• What share of each prime contractor’s revenue is sole-source cost-reimbursement sustainment work — the part most exposed to both open competition and fixed-price default? That ratio, not backlog, is the number that would reprice.
• Which commercial manufacturers have publicly refused defense work over compliance cost, and would they bid now? Contract manufacturers, machine shops, industrial electronics firms and automotive suppliers with spare capacity are the natural entrants.
• If contracting officers can see through primes to suppliers and lend to those suppliers directly, does the prime’s role as the supply chain’s financier disappear, and what is that function worth to the supplier that currently depends on it?
• Defense primes have already fallen a long way from their highs. Is that about program execution and margin misses, or has part of the structural story begun to be priced?
• Who sells the compliance? Government-contract accounting software, government-audit consulting and Cost Accounting Standards advisory are real businesses whose demand this memorandum is explicitly designed to shrink.

Where to look

• Lockheed Martin, Northrop Grumman, RTX, General Dynamics, L3Harris — the incumbents whose competitive position and margin structure the memorandum targets, and the place to test whether any of this is already in the price
• Boeing — the clearest existing case study in what fixed-price default does to a contractor that misjudges the risk it has taken
• tier-two and tier-three defense suppliers — the stated beneficiaries of supplier transparency and direct financing, and the least-covered part of the chain
• commercial manufacturers with defense-adjacent capability, for whom the entry cost is precisely what is being removed
• the Office of Strategic Capital and the Industrial Base Analysis and Sustainment program, as the channel through which money now reaches suppliers without passing through a prime

Thesis check

The document is primary, signed, dated and unusually explicit about what it is trying to break, and the mechanism it names — compliance cost functioning as an entry barrier — is real and has been documented for decades. The weakness is that a memorandum is not a rule: the central piece requires action by the Cost Accounting Standards Board, which sits outside the Department of War, and the defense acquisition system has absorbed and neutralised reform memoranda repeatedly, so implementation risk here is not a caveat at the end but the main question in the middle.

Timing

Memorandum signed September 15, 2026; appendix deadlines and a proposal to the Cost Accounting Standards Board follow

Sources

Department of War, Sep 15 2026 · Fostering One Strong Industrial Base, memorandum and appendix, Sep 15 2026 · Defence Industry Europe, Sep 2026

The Food and Nutrition Administration published eighteen separate Federal Register notices on September 15, 2026, one for each state running a demonstration project under section 17(b) of the Food and Nutrition Act of 2008 that amends the SNAP definition of food for purchase. The states are Arkansas, Florida, Hawaii, Idaho, Indiana, Kansas, Louisiana, Missouri, Montana, Nevada, North Dakota, Ohio, Oklahoma, South Carolina, Texas, Utah, Virginia and Wyoming. Comment periods close October 15, 2026.

Most are already in force, and each state wrote its own list. Indiana and Utah began on January 1, 2026 excluding “soft drinks.” Texas started April 1, 2026 excluding “sweetened drinks and candy,” modified in January 2026 to clarify that naturally sweetened beverages and medical-grade electrolyte drinks stay eligible. Florida began April 20, 2026 excluding “soda, energy drinks, candy, and prepared desserts.” Arkansas began July 1, 2026 excluding “soda, low and no-calorie soda, fruit and vegetable drinks with less than 50% natural juice, other unhealthy drinks, and candy.” South Carolina began August 31, 2026 excluding “candy, energy drinks, soft drinks, and sweetened beverages.”

Ohio and Virginia take effect October 1, 2026. Ohio’s request, approved March 4, 2026, excludes “sugar-sweetened beverages,” and was modified on June 12, 2026 to also exclude “fountain drinks.” Missouri’s project, covering “candy, prepared desserts, and certain unhealthy beverages,” was moved on June 2, 2026 to a February 15, 2027 start.

Trade reporting puts approvals at 23 states covering roughly a third of SNAP recipients, with an estimated $830 million of lost food and beverage sales in 2026, of which about $430 million is soda and $100 million energy drinks. Measured purchase data cited in that reporting shows soda purchase activity falling roughly twice as much in waiver states as in unchanged states, and year-to-date category volume in restricted states running 310 basis points below the US average in soft drinks and 370 basis points below in hard candy.

Share prices verified at the September 15, 2026 close: PepsiCo $135.50, within half a percent of its one-year low of $134.95 and 20.5% below its one-year high; Celsius Holdings $27.63, 57.4% below its high; Monster Beverage $44.63, 10.7% below; Keurig Dr Pepper $31.49, 6.0% below; Coca-Cola $88.71, 3.6% below.

Opportunity

The story as reported is that states are banning sugary soda for food-stamp shoppers, and at that description it is small — a few hundred million dollars against a category worth tens of billions, most of it arguably just a change in which pocket a purchase comes out of. Reading the eighteen notices rather than the coverage gives a different picture, in two ways.

The first is scope. The exclusion lists have drifted well past sugar. Arkansas excludes zero- and low-calorie soda outright, along with any fruit or vegetable drink under 50% juice. Florida and South Carolina exclude energy drinks. Ohio excludes fountain drinks. Zero-sugar is where essentially all the remaining volume growth in carbonated soft drinks sits, energy is the fastest-growing beverage category in the United States, and the fountain is the highest-margin way a soft drink gets sold. A rule aimed at sugar has become a rule aimed at the liquid refreshment aisle and the convenience-store dispenser.

The second is direction of travel. This went from nothing to 23 approved states in roughly eighteen months, no state has reversed, and the notices show states coming back after approval to widen their own lists rather than narrow them — Ohio adding the fountain in June 2026 being the clearest instance.

Hypothesis: the beverage majors are valued on volume stability, and the measured effect — purchase activity falling about twice as fast in restricted states as elsewhere — suggests genuine substitution rather than a change of tender, which would make this a structural volume headwind that compounds state by state rather than a one-off transfer. That is an inference drawn from one set of trade-reported scanner measurements, not an established fact, and at company level the amounts involved are currently small.

How it could play out

Ohio and Virginia switch on October 1, 2026 and Missouri on February 15, 2027, each adding SNAP households to the restricted pool. Retailers across eighteen states carry the cost of item-level eligibility coding across the beverage aisle, and in Ohio across the fountain as well. Measured volume gaps between restricted and unrestricted states widen on a larger sample and begin showing up in company disclosure rather than only in scanner data. If those gaps hold, the political case for the remaining states gets easier to make, and the categories with the fastest growth turn out to be the ones most exposed. If instead households simply pay cash for the same drinks, the effect washes out and the whole thing reduces to a retail compliance cost.

Questions worth asking

• Is the measured decline real substitution or just a change of payment method? SNAP dollars are fungible up to a household’s cash income, so the honest test is total category volume in a restricted state, not SNAP-tendered volume. Everything here turns on that distinction.
• Which companies are most exposed to the categories that got added rather than the ones everyone expected — zero-sugar carbonated soft drinks, energy drinks, and fruit drinks under 50% juice?
• Ohio excludes fountain drinks from October 1, 2026. What share of convenience-store and quick-service gross profit is the fountain, and does anyone disclose it at a level that would let you size this?
• Who bears the compliance cost? Every SNAP retailer in eighteen states needs item-level restriction across millions of barcodes under eighteen different definitions. Is that a burden on small independent grocers and an advantage to chains with central systems?
• Which retailers have the highest SNAP share of sales in these particular states, and does a restricted basket change their mix rather than simply their total?
• Is there a beneficiary? Water, milk, unsweetened drinks and 100% juice remain eligible in every one of these states, and someone sells those.

Where to look

• PepsiCo — carbonated soft drinks plus Gatorade and Rockstar, currently trading within half a percent of its one-year low
• Coca-Cola and Keurig Dr Pepper — the other two large owners of the excluded carbonated categories, including their zero-sugar lines
• Monster Beverage and Celsius Holdings — energy drinks, excluded outright in Florida and South Carolina
• Hershey and Mondelez — candy, excluded in Texas, Florida, Arkansas, South Carolina and, from 2027, Missouri
• Dollar General, Grocery Outlet and convenience-store operators — retailers with high SNAP share of sales in the affected states
• the Federal Register notices themselves, which are the only complete statement of what each individual state actually excludes and from what date

Thesis check

The underlying facts are about as solid as this system ever gets: eighteen dated federal notices naming exactly what each state excludes and when, alongside independently measured purchase declines in the states already running. The real constraint is size — an estimated $830 million of industry-wide lost sales in 2026 is immaterial to Coca-Cola or PepsiCo on its own, so the idea only works if the state count keeps rising and the exclusion lists keep widening, and nothing that has happened so far guarantees either.

Timing

Ohio and Virginia exclusions take effect October 1, 2026; comment periods close October 15, 2026; Missouri follows February 15, 2027

Sources

Federal Register, State of Ohio SNAP Demonstration Project, Sep 15 2026 · Federal Register, State of Texas SNAP Demonstration Project, Sep 15 2026 · Federal Register, State of Arkansas SNAP Demonstration Project, Sep 15 2026 · Food Business News, lost sales from SNAP waivers, 2026 · CNBC, SNAP restrictions and food companies, Jun 20 2026