A magnitude 6.8 earthquake strikes the outskirts of Istanbul. One of Meridian's locations, a commercial office tower worth USD 24M, sustains ground-up structural damage the vulnerability module estimates at 35% of TIV: USD 8.4M of physical destruction.
Three insurers each have a share of this risk. The first has written a policy with a USD 500K deductible and a USD 10M limit, it pays USD 7.9M. The second has written the same risk with a USD 2M deductible and a USD 5M limit, it pays USD 5M. The third has written a policy covering only the structure, not the contents, with a 20% co-insurance clause requiring Meridian to bear a fifth of any loss, it pays USD 4.9M.
Three insurers. One earthquake. One building. One USD 8.4M ground-up loss. Three different insured losses.
The financial module is the last step in the cat model's calculation chain. It takes the ground-up loss produced by the vulnerability module, the total physical damage before any insurance terms are applied, and transforms it into the loss the insurance policy actually covers. That transformation is governed by the policy conditions: the specific contractual terms that define what the insurer pays, under what circumstances, and up to what limit.
This step is what makes cat modelling useful for insurers rather than just for engineers. An engineering model can estimate how much a building is damaged by a given force. Only a cat model with a properly configured financial module can translate that damage estimate into the number the insurer cares about: how much of that damage falls within the insurance contract and therefore becomes the insurer's liability.
The financial module does not change the physical reality of the loss. If the building sustains USD 8.4M of damage, USD 8.4M of damage occurred regardless of what any insurance policy says. What changes is the allocation of that loss between the insured (who bears the deductible and anything above the limit) and the insurer (who bears the covered portion between the deductible and the limit). The financial module calculates each party's share.
Every loss figure a cat model outputs exists at a specific point in this transformation. Ground-up loss is the raw physical damage. Gross loss is after the primary policy conditions are applied. Net loss is after reinsurance recoveries are also applied. When an underwriter asks "what's the model saying?" they need to know which of these three they are looking at, the same event produces three materially different numbers, and confusing them leads to mispriced risk or miscalculated reinsurance recovery.
Policy conditions are the contractual terms that define the coverage. They tell the financial module exactly how to translate physical damage into insured loss. The four most important conditions for cat modelling purposes are deductibles, limits, coverage scope, and participation.
The amount the insured must bear before the insurance pays anything. Acts as a floor below which the insurer has no liability. A higher deductible reduces the insurer's expected loss, the insured absorbs all losses below it.
The maximum the insurer will pay under any single loss or in any policy period. Acts as a ceiling above which the insured bears the excess themselves. The policy limit caps the insurer's liability regardless of how large the loss grows.
What the policy actually covers - which perils, which types of loss (structure vs contents vs business interruption), and which geographic locations. A policy covering wind but not flood pays nothing for a flood loss, regardless of its limit. Coverage scope is the first filter the financial module applies.
Some policies require the insured to bear a defined percentage of every loss, typically because the insured is self-insuring part of their own risk. A 20% co-insurance clause means the insurer pays 80% of the covered loss above the deductible, regardless of how large it grows.
The financial module applies these conditions in sequence for every event in the stochastic catalogue. For each of the tens of thousands of simulated events, it takes the ground-up loss the vulnerability module produced, subtracts the deductible, caps at the limit, applies coverage scope filters, and calculates the insured's participation share. The result, applied across every event in the catalogue, produces the gross loss distribution from which the EP curve and AAL are derived.
Cat model output exists at three distinct levels of the loss calculation chain. Each level strips away another layer of financial transformation to show the loss from a different perspective. Knowing which level you are looking at is not optional, the same event produces three meaningfully different numbers and conflating them is one of the most consequential errors a practitioner can make.
The distinction matters differently for different stakeholders. An underwriter pricing a single risk thinks primarily in gross loss, what is Halcyon's liability on this policy? A reinsurance actuary structuring the Cat XL programme thinks in the difference between gross and net, what recovery does Halcyon expect from its reinsurers in a major event? The capital model and CRO think in net loss, what does Halcyon actually need to hold capital against after all reinsurance recoveries? The financial module is what makes all three calculations possible from the same underlying ground-up loss.
The financial module can only apply policy conditions it has been told about. If the exposure file does not include the deductible amount, the module cannot apply a deductible, it defaults to zero, treating every dollar of ground-up loss as insured. If the limit is missing, the module cannot cap the loss. If coverage scope is not specified, it applies the broadest available coverage. Every missing policy condition biases the gross loss estimate upward.
This creates a systematic relationship between exposure data quality and loss output. The three preceding lessons established that exposure data quality affects the hazard assignment (geocoding), the damage ratio (COPE), and the financial multiplier (TIV). The financial module adds a fourth dimension: the quality of the policy condition inputs determines how accurately the ground-up loss is converted to insured loss. A model run with perfectly geocoded locations, confirmed COPE, and accurate TIVs will still produce unreliable gross loss estimates if the policy structure has not been configured correctly.
By the end of Module 02, the pattern should be clear. The cat model chain, hazard → vulnerability → financial, has an exposure data input at every stage. Geocoding quality determines hazard. COPE quality determines vulnerability. TIV and policy conditions determine the financial output. A model run is only as reliable as its weakest data input, and that weakest input can sit at any of the three stages. Checking data quality means checking all three, not just the location coordinates.
Deductibles affect the insured loss at every event size, not just large ones. A USD 500K deductible eliminates the insurer's liability for any ground-up loss below USD 500K, which may be a significant fraction of all events in a high-frequency, moderate-severity peril like severe convective storm. The financial module affects the AAL (expected annual loss) materially because it shapes the entire loss distribution, not just the tail.
They are fundamentally different quantities. Gross loss is what the primary insurer owes the policyholder. Net loss is what the primary insurer retains after its own reinsurance pays. For a well-reinsured syndicate, net loss in a major event could be a fraction of gross loss. Presenting gross loss as the retained exposure overstates the capital requirement; presenting net loss to the policyholder understates their coverage.
Only if the ground-up loss exceeds the deductible and is large enough to benefit from the higher limit. For small events where ground-up loss is below the deductible, both a USD 5M and a USD 50M limit produce zero insured loss. The financial module produces the same outcome for both policies at low severity levels. The limit only matters when the loss is large enough to approach or exceed it.
The financial module applies deterministic arithmetic to the ground-up loss. But if the policy conditions fed into it are wrong or incomplete, it introduces systematic bias into every output it produces. A misconfigured deductible does not introduce statistical uncertainty, it introduces a directional error that shifts the entire gross loss distribution. This is worse than statistical uncertainty because it is consistent and undetectable without knowing the correct policy terms.
The fourth cat model component. Takes the ground-up loss from the vulnerability module and applies policy conditions to produce the insured loss. Operates on every event in the stochastic catalogue to produce the gross loss distribution.
Total physical damage to the insured property before any policy terms are applied. The output of the vulnerability module and the input to the financial module. Represents the full economic scale of the damage regardless of insurance structure.
The loss the primary insurer owes the policyholder after applying primary policy conditions, deductible, limit, coverage scope, and co-insurance. Before the insurer's own reinsurance protections are applied.
The loss the primary insurer retains after all reinsurance recoveries, Cat XL and other protections, have been applied to the gross loss. The net loss is what the insurer's capital must absorb. The relevant figure for solvency capital calculations.
The amount the insured bears before the insurance policy pays. Acts as a floor on the insurer's liability, every dollar of ground-up loss below the deductible falls entirely on the insured. Reduces the insurer's expected loss across the full event distribution.
The maximum amount the insurer will pay under a policy. Acts as a ceiling on the insurer's liability, losses above the limit fall on the insured unless additional coverage exists. The insurer's maximum single-event exposure at a location is the limit, not the TIV.
The contractual definition of what the policy covers, which perils, which loss types, which locations. The first filter the financial module applies. A flood loss on a wind-only policy produces zero insured loss regardless of the ground-up damage.
The full set of contractual terms that govern the financial module's transformation of ground-up loss into insured loss, deductible, limit, coverage scope, co-insurance, and any special conditions or sublimits. Configured from the policy slip, not the SOV.
Five questions · 4 of 5 correct (80%) to pass