Validation with the depth of a builder.
Credit, cash-flow, forecasting, and capital models drive financial, budgetary, and policy decisions — and every model carries risk in its data, assumptions, design, and use. As reliance on models grows, so does the need to govern them across their full lifecycle.
Our hands-on experience building these models lets our validation reach the substance of methodology and calibration — not just documentation. ALT helps agencies and institutions govern model risk through validation, monitoring, and oversight aligned with current interagency guidance (SR 26-2). We work as a reviewer or a development partner on a given model, never both — so independence is preserved on every engagement.
What strong model governance delivers
Model governance across the lifecycle
Model validation
Validate credit, cash-flow, PD/LGD, prepayment, forecasting, and capital models against intended use.
Challenger & benchmark models
Build independent challenger models to reproduce results, stress conclusions, and expose weaknesses.
Ongoing performance monitoring
Design back-testing, drift, and stability monitoring with thresholds that trigger timely re-review.
Model development & enhancement
Develop, recalibrate, and document models as a build partner, engaged separately from validation.
Governance, policy & documentation
Strengthen model risk policy, inventory, roles, tiering, and documentation across the enterprise.
Findings & remediation advisory
Prioritize findings by severity and advise on practical, defensible remediation and re-review.
How a model earns confidence
Scope the model & materiality
Clarify purpose, materiality, intended use, and the tier and standard of review the model requires.
Examine soundness, data & assumptions
Review conceptual design, data quality and lineage, assumptions, calibration, and implementation.
Build the challenger & test
Reproduce results, build an independent challenger, and run back-testing, benchmarking, and sensitivity analysis.
Document, report & advise
Deliver findings, severity-ranked recommendations, and remediation guidance owners can act on.
Questions we help model owners answer
- Is this model conceptually sound, well-calibrated, and fit for its intended use?
- Do the results hold up when an independent challenger reproduces and stresses them?
- Is model performance drifting, and would we detect it before it affects decisions?
- Are our governance, documentation, and controls strong enough to withstand scrutiny?
Connected ALT Solutions services
Need confidence in a model — or help building one that will hold up?
ALT helps agencies and institutions validate, challenge, monitor, and govern the models behind high-consequence decisions.