Underwriting & pricing analytics for installment and cards

Examine how eligibility rules, credit decisions, pricing, and discretionary actions affect outcomes across defined consumer-lending populations.

Map the decisions separately

Installment loans and revolving card products have different journeys and outcome measures. Separate application eligibility, approval, credit limits, pricing, conditions, and subsequent account actions rather than combining them into a single score. Identify which decisions are rule based, model supported, or discretionary. The analysis should reflect the institution’s actual decision process, including channels and handoffs that are not visible in a central model output.

Define comparable populations

Create documented product and time-period definitions before comparing outcomes. Record policy versions, campaign eligibility, channel differences, missing information, and material changes in the application mix. Retain the rationale for analytical variables and the limitations of each comparison. An apparent pricing difference can require additional investigation; it should not be explained away through unsupported assumptions or treated as a complete finding without review.

Join pricing to the decision record

Link the offered and accepted terms to the application, the governing policy, relevant model output, and any approved adjustment. Distinguish the initial offer from later negotiations or changes. Capture override reasons in a controlled vocabulary while retaining enough narrative to understand unusual cases. Reconcile the analytical fields to the authoritative lending record and document how unavailable or conflicting values were handled.

Use analytics to direct review

Present patterns with sample sizes, missing-data information, and the decision context needed by lending, compliance, and risk teams. Define who reviews an exception and when an expanded investigation is warranted. Compare repeated periods using stable definitions and disclose method changes. Cicrim helps connect consumer-lending data, analytical workflows, and evidence routines so the institution can make reviewable decisions without claiming that one metric establishes the whole customer outcome.

Working-session checklist

• Product and decision-stage scope
• Policy and pricing versions
• Outcome and adjustment definitions
• Comparable-population rationale
• Review and escalation ownership