Skip to content

Extending the Common Domain Model for Property and Physical Risk #3432

Closed

Description

@eteridvalishvili

Presenting: Kelly, David (Bip Group) and McGoveran, Fearghal (Bip Group) and Johnny Mattimore (BIP Group)

Materials:

CDM - Extending the Common Domain Model for Property and Physical Risk

The CDM has gained substantial traction within the banking community by delivering meaningful workflow and cost benefits for adopters. Given the material increase in existing and future exposure that banks have to physical risk events on property assets, the Physical Risk Programme of FINOS, under OS-Climate, is extending the CDM to provide comprehensive foundational data fields to support end-to-end workflows for the integration of physical hazards from flood to fire, wind and beyond.

This CDM extension will allow banks to reduce the time and costs to deliver internal and external reporting, deep analytical insights for transaction and portfolio review and management, as well as advanced risk simulation for scenario and stress testing. This will ultimately extend to the ability to transfer risk using the concept of Physical Risk Swaps for hedging and wider financial product creation to meet different risk appetites in the financial ecosystem.

To achieve this, we are taking the CDM and extending it as a "foundation data model" for banks. We are then integrating it with the valuable components that feed solutions for existing Catastrophe Risk used widely for insurance and repurposing them with new components, including NVIDIA's AI Platform, to create bank specific solutions.

This will be a unique Open Source solution at speed, scale and low cost to address current and future Physical Risks faced by banks.

Activity

  1. moved this to In Review in CDM CRWGon Feb 18, 2025
  2. added
    complexMajor change; should be broken-down into smaller units; Steering WG approval recommended
    on Feb 18, 2025
  3. moved this from In Review to Reviewed in CDM CRWGon Feb 18, 2025
  4. dshoneisda commented on Mar 3, 2025

    @dshoneisda
    Contributor

    Please raise any new WG proposals to the SWG via @eteridvalishvili
    We will need a proposed scope and chair to ratify any proposal, though this feels different enough to warrant it's own domain WG. Note that a senior WG would still need to approve contributions once the domain WG is satisfied with any model contribution proposal

  5. chrisisla commented on Mar 4, 2025

    @chrisisla
    Contributor

    Thank you for providing the materials that you presented at the last CDM CRWG. Could you clarify whether you are proposing to add content to the core CDM or create an extension to it? Or both?

    From what I can tell it looks like the end result of your proposed changes is to create a new physical risk swap. Would this be a new product type in the CDM?

    The proposal refers to a lot of additional data to support the new swap type e.g. details of flood gauges, property, insurance. I have a concern with this as there looks to be a lot of reference data that has little overlap with the other financial products modelled in the core CDM. Is the intention to add this to the core CDM, or manage this in an extension?

    It looks like the proposal will now be brought to the CDM Steering Working Group (SWG) for discussion. At that meeting it would be great to get some clarity on what changes are being proposed to the core CDM, and what would be held in an extension.

    Thanks!

  6. moved this from Open to Scheduled in CDM Steering WGon Mar 10, 2025
  7. iansloyan commented on Mar 11, 2025

    @iansloyan
    Contributor

    Thank you for providing the materials that you presented at the last CDM CRWG. Could you clarify whether you are proposing to add content to the core CDM or create an extension to it? Or both?

    From what I can tell it looks like the end result of your proposed changes is to create a new physical risk swap. Would this be a new product type in the CDM?

    The proposal refers to a lot of additional data to support the new swap type e.g. details of flood gauges, property, insurance. I have a concern with this as there looks to be a lot of reference data that has little overlap with the other financial products modelled in the core CDM. Is the intention to add this to the core CDM, or manage this in an extension?

    It looks like the proposal will now be brought to the CDM Steering Working Group (SWG) for discussion. At that meeting it would be great to get some clarity on what changes are being proposed to the core CDM, and what would be held in an extension.

    Thanks!

    Hey Chris,

    Perhaps as I have been close to this development I can field these questions in advance of the Steering WG (also as I can only attend the first 15 mins today!) and give the necessary CDM perspective.

    Initially it would not be an extension to CDM, if in time preferable for governance and technical reasons (and if a roaring success where a new cohort want to help evolve further quickly), it could be made a separate extension model which inherits from CDM - but to keep things simple developing the property type and related risk attributes below it (and the reference data enums etc) in CDM core makes most sense.
    It should be very much contained but we want to see the good practices and basic data types of CDM inherited directly in this property and physical risk model. (A bit like the collateral specification.) Property is such a fundamental asset class that I think this expansion makes sense. Loan already exists, and Collateral might be another type where property is referenced in as an attribute.

    While true, the long term evolution of the project to define these risk attributes should end up in a physical risk swap, that product would be modelled on same structures as other ISDA derivative products (most akin to CDS).
    Remembering that strictly there are no product types in CDM, there are only payouts (and qualification functions that can put multiple labels on trades in a object called ProductTaxonomy).
    To model a physical risk swap would probably require use of a interestRatePayout and a CDS-like payout - so there could be a new payout, but I think generalising CreditDefault to capture payouts based on other discrete events/observations or using the PerformancePayout for binary observations would suffice. This is a secondary goal though. Modelling the risk and property features is the primary goal and that should be quite contained in an attribute at the end of a path like Loan in instrument.
    So the headline is probably that we are creating a type called Property with lots of detail to it related to risk....then we can discuss if it should be attached as a new instrument, or only as underlying collateral to Loan. (That is when wider discussion with the Maintainers and CDM community can take place.)

  8. chrisisla commented on Mar 11, 2025

    @chrisisla
    Contributor

    Thanks for the reply @iansloyan

    I actually disagree with the "get it into core CDM first, then take it out as extension if necessary" approach, I think it should be the other way around, i.e. prototype/develop externally first and then bring in the "core" parts if they really are core, especially since the majority of the changes look to be reference data and there is a project currently underway to specifically remove reference data from the core model.

    So the headline is probably that we are creating a type called Property with lots of detail to it related to risk....then we can discuss if it should be attached as a new instrument, or only as underlying collateral to Loan.

    A new Property type on its own, as a new Asset type, would make sense, but on first glance I don't think all the supporting reference data should be in the core. I'm sure we can agree that we don't want the model to become so large that it loses sight of its ability to support its original use cases.

    (That is when wider discussion with the Maintainers and CDM community can take place.)

    I believe that the wider discussion should be done before adding anything to the core model - which I understood was the reason for the presentation at the CRWG and at today's SWG - so that the maintainers can comment before too much work is undertaken.

  9. iansloyan commented on Mar 11, 2025

    @iansloyan
    Contributor

    Apologies if I am misrepresenting the intentions here. To be clear - everything is developed and prototyped as an extension to the model firstly (effectively a branch) before being committed to the CDM in a version release via a maintainer approved PR (and as I said the contained nature of this should allow it to be done as one PR after full agreement rather than many changes). Nothing will change here in that sense. But this work does not require a new "model" like say DRR. It should all be part of the CDM project. Just like the detailed data model for legal agreements, specification of securities in collateralSpecification or detailed date attributes in payout structures. These are risk features of a key asset class.

    If there is a concern about ref data, classification of the attributes which will form part of this work as ref data or fundamental types/attributes is for discussion in the working group. And as you suggest keeping reference data external to the model can only be achieved when that other functionality is delivered - and I would hope that work will also deliver some principles for what is ref data and what is not (as in the abstract it seems hard to define - maybe like some things it is hard to define but you know it when you see it!)

  10. chrisisla commented on Mar 11, 2025

    @chrisisla
    Contributor

    Hi @iansloyan
    This was covered a bit in the SWG earlier. It was stated that the scope of the changes would be derived over the first few WG sessions, and that is really where my question was coming from, what is the actual scope of the work here? You're obviously much closer to this than I am, all I have seen is the deck attached to this Issue which does not really provide enough clarity over what the changes to CDM core will be and their overall impact on the model. Over the coming weeks it sounds like this will be defined which is what I wanted to confirm.
    Thanks!

  11. moved this from Current to Approved in CDM Steering WGon Mar 18, 2025
  12. CDM-ReleaseManagement-AP commented on Jul 1, 2025

    @CDM-ReleaseManagement-AP
    Contributor

    Presented at the CDM Contribution Review Working Group - July 1st, 2025, removed from CRWG Projects board

  13. lolabeis commented on Jul 1, 2025

    @lolabeis
    Contributor

    @jmattimore (as WG chair):

    It was suggested at the CRWG that this initial issue, which led to the creation of the WG, be closed, and new issue(s) be raised in the Physical Risk WG as this WG's units of work / contributions start to materialise.

  14. moved this from Approved to Closed in CDM Steering WGon Jul 31, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    complexMajor change; should be broken-down into smaller units; Steering WG approval recommended

    Projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions