Published October 2026.
Direct Answer: Catastrophe model output is not a price tag—it is an input to technical premium, retention, and limit choices that underwriters then reconcile with capacity, appetite, and market competition. When a carrier says “the model says $40M,” your job is to trace that number through peril scope, return period, basis (ground-up versus gross), loadings, and reinsurance so you know what rate, deductible, and attachment actually follow—and when a documented override is in play.
You finished renewal prep with modeled average annual loss near $2.1M and a 1-in-100 aggregate near $40M. The indication lands 18% above last year’s rate on line and pushes a higher wind deductible. Both sides cite “the model.” This article walks the last mile: modeled loss to rate, retentions, technical versus charged premium, overrides, and questions to ask before you sign.
What the carrier’s number usually is—and what it is not
Before you argue rate, align on definitions. Most carrier conversations mix several distinct outputs:
- Modeled average annual loss (AAL)—expected loss per year before reinsurance, often ground-up or subject to a stated gross/net convention.
- Occurrence or aggregate metrics—for example a 1-in-100 occurrence loss or a 1-in-250 aggregate, each answering a different capital question.
- Technical premium—AAL plus loadings for expenses, cost of capital, and volatility, sometimes scaled to a target return period.
- Charged premium—what appears on the quote after competition, portfolio strategy, and underwriting judgment.
A broker slide that shows $40M is rarely “your premium.” It is more often a tail metric the carrier uses for capacity, reinsurance purchase, or regulatory reporting—not an invoice line item. For how models are built, see the catastrophe modeling complete guide; this piece stays on the decision layer after those outputs exist.
From modeled loss to insurance rate: a worked path
Underwriters start from expected loss, then apply loadings. Illustrative arithmetic keeps the logic visible when carriers will not share every factor.
Step 1: Ground-up AAL to subject AAL
Take a $500M total-insured-value schedule across ten coastal properties. The cat platform returns ground-up hurricane AAL of $1.8M. Your policy has a $100,000 per-occurrence deductible and a $25M per-occurrence limit. The model’s “subject” AAL—the portion that would hit the policy after deductibles and limits—might fall to roughly $1.4M after those terms are applied in the engine. Always confirm whether the carrier quoted ground-up or subject; mixing them is a common renewal dispute.
Step 2: AAL to technical rate on line
Technical premium often follows a form like:
Technical premium ≈ (Subject AAL ÷ limit) × load factor
Assume subject AAL $1.4M, limit $25M, load factor 1.65 (expense, cost of capital, and catastrophe volatility margin combined). Subject loss ratio on limit: $1.4M ÷ $25M = 5.6%. Technical rate on line ≈ 5.6% × 1.65 ≈ 9.2%. On a $25M limit, technical premium ≈ $2.3M before reinsurance credit.
Reinsurance changes the story: excess-of-loss cover can cut net expected loss while the gross quote still reflects gross model loss in tight markets. See reinsurance treaty structures explained for how layers interact with modeled aggregates.
Step 3: Technical versus charged premium
Charged premium adjusts technical premium for competition, relationship, and portfolio mix. A carrier may run 9.2% technical and quote 11.5% charged when cat capacity is scarce—or 7.8% to retain a strategic account. Reconstruct technical premium from shared assumptions rather than debating one percentage in isolation.
Retention and limit: translating tail output into structure
Model output drives not only rate but how much risk you keep. Underwriters watch tail metrics (100-year, 250-year) alongside AAL when setting deductibles, self-insured retentions, and policy limits.
Deductibles from modeled frequency and severity
Suppose wind deductibles are expressed as 2% of TIV per location, minimum $250,000. Modeled results show that lowering the minimum to $100,000 adds roughly $180,000 to subject AAL on your schedule because more small and medium events pierce the retention. The underwriter proposes raising the minimum to $500,000, which the model shows reduces subject AAL by about $220,000 but leaves you retaining more ground-up loss on moderate events. The trade is arithmetic plus appetite: cash liquidity versus premium savings.
Limits and attachment from tail losses
If the 1-in-100 occurrence loss ground-up is $85M and you buy $50M primary with a $50M excess layer attaching at $50M, the model’s gross loss at that return period informs whether the tower is credible or cosmetic. Portfolio managers also watch accumulation across locations; when multiple properties correlate in one event, location-level limits understate true exposure. For accumulation, PML, and reinsurance alignment, catastrophe portfolio management covers the upstream view—here, the underwriting question is whether your requested limit matches the modeled loss at the return period the carrier actually capitalizes to.
Technical premium versus market price at renewal
Market price is what clears in the placement window. Technical premium is what the carrier’s internal models say is required for a target return. In soft markets, charged rates sit below technical for many accounts; in constrained cat markets, charged rates exceed technical for peak zones because supply caps bite before math does.
Document three numbers in your renewal file when possible: modeled subject AAL, implied technical rate on line, and quoted charged rate. The spread between technical and charged is information. Persistent charged-above-technical may signal sector stress or account-specific concerns (concentration, loss history, data quality flags). Charged-below-technical may signal competition—or an underwriter who expects remedial terms later if loss experience deteriorates.
Modeling costs affect how much independent checking you can afford; see catastrophe modeling costs and fees. A lighter run on the same exposure snapshot helps test whether the carrier’s subject AAL is in the same ballpark.
When underwriters override the model—and how they should document it
Models are tools, not statutes. Overrides are normal when empirical data or engineering judgment materially disagrees with catalog assumptions—provided they are scoped, authorized, and recorded.
Common override triggers
- Site-specific mitigation—hurricane strapping, flood barriers, or wildfire defensible space not fully captured in geocoding or building attributes.
- Book versus file mismatch—submission TIV or construction class differs from modeled values after inspection.
- Peril scope differences—flood excluded on policy but still appearing in a blended “nat cat” dashboard export.
- Recent loss experience—credible local events that stress-test vulnerability beyond average years.
- Portfolio strategy—temporary appetite expansion or contraction not encoded in vendor defaults.
Overrides should state direction, magnitude, peril, dates, and approver. Weak documentation surfaces at audit and claim time. Ask for a summary when quoted terms move materially from a transparent technical build—especially deductible or sublimit changes tied to “model update.”
Questions to ask when the carrier cites “the model”
Polite, specific questions beat generic pushback. Use them in broker strategy calls and underwriter meetings:
- Which metric are you citing—AAL, occurrence loss at 100 years, aggregate at 250 years—and on what basis (ground-up, gross, net of reinsurance)?
- Which event set and version date were used, and did peril scope match our policy (wind-only versus full nat cat)?
- What load factor converts AAL to technical premium for our limit and layer?
- How much of the rate change is model movement versus loading, expense, or market adjustment?
- What retention change produces how much modeled premium delta on our schedule?
- Was any underwriting override applied? If yes, what attribute or experience drove it?
When one peril dominates your schedule, hurricane, earthquake, and wildfire peril analysis adds context; you still only need consistent definitions to negotiate rate and retention from the carrier’s authorized output.
How buyers should use model output at renewal
Treat your model export and the carrier’s model as two witnesses, not one verdict. Align inputs first: same locations, values, construction, deductibles, and limits. Then compare subject AAL and one agreed tail metric. If gaps persist, split “data disagreement” from “terms disagreement.” Data gaps get fixed with surveys and engineering; terms gaps get structured with deductibles, limits, and alternative markets.
Build a one-page renewal logic map: subject AAL, implied technical rate, quoted rate, retention deltas, and open override questions. Share it early so marketing targets carriers whose appetite matches your modeled profile.
When catalogs refresh, ask what changed on your schedule—frequency, severity, or indexing—not only the headline percent change. See climate risk pricing and catastrophe model updates for background; at renewal, tie term movements to pieces you can verify.
Present ranges from your own vendor run and ask the carrier to reconcile to their authorized platform. How major vendors differ is covered in RMS, AIR, and Verisk in catastrophe modeling; here the objective is a decision-ready bridge to bound terms.
Closing the gap: dashboard to deal
The model says $40M at a stated return period; your job is to translate that into whether $2.3M technical premium for a $25M limit feels fair at a 9.2% rate on line, whether a $500,000 wind minimum is worth $220,000 of modeled premium relief, and whether overrides are documented when the quote diverges from transparent math. Carriers that explain that chain earn faster binds; buyers who understand it avoid signing structures priced on metrics they never agreed to. That is the last mile between the dashboard and the deal—and it is where catastrophe modeling pays off in dollars, not slides.
FAQ
What is the difference between technical premium and charged premium?
Technical premium is the amount an insurer’s pricing framework says is needed to cover modeled expected loss plus planned loadings for expenses, cost of capital, and catastrophe volatility. Charged premium is what the insurer actually quotes or binds after competitive pressure, account strategy, and underwriting judgment. The two often differ; understanding both lets you see whether a renewal move is driven by model loss change or by market and discretion.
How does modeled loss become an insurance rate?
Underwriters usually convert subject average annual loss into a loss ratio by dividing AAL by the policy limit, then multiply by a load factor that reflects expenses and return requirements. That yields a technical rate on line; charged rate adjusts from there. Always confirm whether AAL is ground-up or subject to your deductibles and limits before comparing rates year over year.
How are retentions set from model output?
Carriers simulate how deductibles, self-insured retentions, and co-insurance change the loss that attaches to the policy. Each retention option produces a different subject AAL and tail loss, which maps to a different premium. Underwriters balance modeled premium savings against your liquidity and the frequency of events that will pierce higher retentions.
When do underwriters override the model?
Overrides are common when verified mitigation, corrected exposure data, peril scope mismatches, or credible loss experience are not fully reflected in the vendor run. Effective overrides are directional, quantified where possible, approved at the right authority level, and documented with peril and date context. Buyers should ask for a plain-language summary when terms diverge materially from a transparent technical build.
What should I ask when the carrier cites the model?
Ask which metric and return period they cite, ground-up versus subject basis, event set and version date, load factors used for technical premium, and how much of the change is model versus market adjustment. Ask whether overrides were applied and how a proposed deductible or limit change shifts modeled premium on your schedule. Request consistency with the policy’s peril scope and your submitted values.
How should a buyer use model output at renewal?
Reconcile inputs with the carrier, compare subject AAL and at least one tail metric on agreed terms, and separate data fixes from structure negotiations. Build a short logic map from modeled loss to implied technical rate to quoted premium and retention choices. Use independent or broker runs to sanity-check subject AAL, and press for documentation when charged premium diverges from explainable technical premium.