Interest Rate Risk (IRRBB) Management demo
Agents back-test every key assumption against what members did, put a number on each gap, and send changes through a validator and ALCO 1:29
What the demo shows
ALM teams govern IRR assumptions, back-test models and document NII and EVE limits
An Extract Reconciler pulls the quarter's core and loan servicing extracts and ties them to the general ledger. A Back-test agent measures each key assumption on what members actually did and explains every gap with cited causes. A Sensitivity agent re-runs NEV and NII in the ALM model with the observed value, alone and with every other open change, and a Limits Watch agent reads the result against board policy and the NCUA NEV Supervisory Test. A Governance agent drafts the change memo. A second-line Challenge agent, owned by Model Validation, recomputes the number on its own. The ALM Manager proposes, the validator accepts or challenges, ALCO votes, and only then is the new assumption set released to the model.
Transcript
- Interest rate risk management keeps a credit union's assumptions honest.
- NCUA expects supportable assumptions, changed on observed evidence.
- At Norrell Federal Credit Union, twenty-eight assumptions are promises against what members did.
- Regular share decay was assumed at fourteen percent; members ran off at twenty-two point six.
- The Back-test agent cites the core: forty-one thousand eight hundred accounts closed or drained.
- Re-run at plus three hundred, NEV drops from nine point one to eight point two.
- The Governance agent drafts the memo, every figure cited, and Nadia Ferrante submits it.
- The validator's own agent recomputes it; he accepts, with a January re-test.
- Until ALCO approves, the model keeps fourteen percent.
- Every step lands in the audit trail, with one NCUA examiner pack.
- Every assumption, back-tested every quarter, with its impact.
- Built on Botminds: agents, data, screens and governance on one platform.
The narration as written for the film. Captions are on the player.