Cross-Domain
Transmission Channels

A practical model for tracing how a single shock moves between the seven Abgalis risk domains — and what that means for capital, underwriting and disclosure.

The siloed risk model is incomplete

Open the risk register of any large insurer and you will find seven or eight headline categories — insurance, credit, market, operational, group, conduct, strategic, climate — each with a sub-taxonomy, an owner, and a quarterly committee. Each is treated, in the modelling layer, as a separate distribution. Aggregation happens through a correlation matrix calibrated to historical co-movement.

This works — until it doesn't. The patterns insurers have found difficult to explain in the past decade share a common structure. A peril event arrives in one domain; the loss it triggers there is modelled and reserved; but the second-order consequences in adjacent domains arrive months later, with no reservation, no capital allocation, and frequently no owner. The insurer accounts for the shock as an insurance loss while the same shock is generating a credit migration in the bond portfolio, an operational continuity failure in the reinsurance recoverable, and a regulatory letter that constrains next year's underwriting.

Seven domains, restated

ABGALIS organises the enterprise into seven domains — chosen because each has a coherent transmission identity:

What a transmission channel is

A transmission channel is the route by which a shock in one domain becomes loss, capital pressure, or constraint in another. Channels have three observable properties:

• Direction — channels are not symmetric. Climate physical shocks transmit fast to insurance and operational, slower to credit and regulatory.
• Latency — some channels move within days (a hurricane creates an insurance loss and a programme drawdown almost simultaneously). Others take quarters.
• Amplification or attenuation — some channels deliver less than 100% of the shock, others more, depending on concentration, leverage and feedback.

Why correlation matrices keep failing

Three structural reasons. Calibration regime — matrices learn from quiescent periods, then aggregate as if the same coupling held in the tail. No direction or latency — a symmetric coefficient cannot express timing. Conflation of channel with co-cause — two domains can be correlated because a third factor causes both, or because one channel transmits the shock directly. The risk-management response should be different in each case; the matrix treats them identically.

What the Abgalis Engine adds

The Abgalis Engine is the cross-domain intelligence layer that powers ICRIP. It treats transmission channels as first-class entities — each with identity, direction, latency, and a structural transmission characteristic that can be calibrated, monitored and stress-tested. The Engine doesn't replace the SCR; it constrains the dependency module of the SCR architecture by ruling out diversification credit between domains where the apparent independence is an artefact of the calibration window.

Output is a structured account of where a shock travels and when, which existing capital infrastructure consumes in three ways:

• Stress-test design — single shock, traced with explicit latency across all seven domains, output as a multi-domain time-indexed cash-flow.
• Reverse stress — channel structure inverts cleanly. Given a binding capital outcome, identify the minimum-magnitude shocks that produce it.
• Disclosure narrative — ISSB, CSRD, TNFD, SS5/25, DORA increasingly ask insurers to describe how risks transmit. A channel model produces that narrative as a by-product.

What this is not

• Not a replacement for the SCR standard formula or an internal model — both remain.
• Not a forecasting tool — the Engine traces what happens once a shock has originated.
• Not a black box — every channel can be inspected, calibrated against the firm's data, challenged, and overridden in the firm's ORSA narrative.

The internal mathematical methodology is patent-pending and reserved to ABGALIS. This page describes the surface — the channels, their properties, and their use in capital and disclosure work.