Credit Portfolio Decision Diagnostics

Who This Is Relevant For

Risk leaders are expected to justify model behaviour — not just monitor metrics.Decisions around PD calibration, thresholds and policy changes are routinely reviewed by model risk, internal audit and committees. The challenge is rarely the availability of reports, but understanding whether observed portfolio outcomes match intent.Situations we are typically asked to review:

• Calibration drift identified only after portfolio performance changes
• Threshold or policy adjustments made without quantified approval and loss impact
• Monitoring spread across multiple reports without a single decision narrative

Our Typical Engagements

Model Decision Review
Independent evaluation of whether model outputs translate into intended approvals and losses.

Calibration & Stability Assessment
Identify drift, segment instability and rebuild vs recalibrate decisions.Threshold & Policy Impact Analysis
Quantify approval, loss and ECL trade-offs before production changes.Governance & Monitoring Structure
Create a defendable framework for ongoing model oversight and committee documentation.

How a Typical Review Engagement Works

Our reviews are structured and lightweight. Most of the work is performed offline using extracts rather than requiring system integrationPortfolio Understanding & Data Review
We review score outputs, performance data and segmentation to understand how the model is being used in practice and confirm data suitability.Model Behavior Assessment
We evaluate calibration, discrimination and stability — both overall and across key segments — to determine whether outcomes align with expectations.Decision & Strategy Analysis
We quantify how thresholds or policy adjustments affect approvals, losses and ECL under realistic operating constraints.Findings & Governance Summary
You receive a concise, committee-ready summary explaining observations, implications and whether recalibration, rebuild or no action is appropriate.