Retail Credit Application Scorecard

Scores personal loan and card applications, with reason codes.

Strictly ConfidentialTier 1 modelCredit decisionBanking secrecy+1High riskScoring · 4.1.0 · retail-credit-scorecardOwner Credit Risk

Test in Sandbox

46 38K/30d 58

AUC (out-of-time)

0.781

Gini

0.562

0.428

PSI (score, last quarter)

0.04

1 more metric

Adverse impact ratio (gender)

0.96

About this model

Application scorecard for personal loans and credit cards: a points-based logistic scorecard over binned bureau, income and relationship attributes, producing a score, a probability of default over 12 months, a risk grade and up to four adverse-action reason codes. Deliberately kept as a transparent scorecard so every decision can be explained to the customer and to the regulator.

Intended use

Input to the retail credit decision engine alongside policy rules and underwriter judgement. Tier 1 model: any change in cut-offs goes through Credit Risk committee. Customers declined receive the reason codes in plain language.

Training data lineage

Rebuilt from the SAS Enterprise Miner scorecard that ran retail originations since 2018. Developed on loan-applications-history (through-the-door applications 2021 to 2025, with 12-month performance) and relationship features from customer-360-mart in BigQuery. Binning, WoE and logistic fit run in Vertex AI Pipelines; independently validated by Model Risk Management, with fairness testing across gender, age band and ethnicity proxies. Annual revalidation; champion-challenger against a gradient-boosted challenger.

Limitations & bias notes

Thin-file applicants (no bureau history) fall back to a separate policy grid. Reject inference is based on bureau outcomes and carries uncertainty. PSI is monitored monthly; a value above 0.25 on the score triggers a review by Model Risk Management.

#credit-risk#scorecard#application-scoring#explainability#fairness#tier-1#rebuilt-from-sas

Ownership and sensitivity

Owned by

Group Risk Management Credit Risk

Accountable owner: Rajesh Kumar

Who approves access

Owner, Credit Risk, Rajesh Kumar
Data Governance Office, Ahmad Faizal
Model Risk Management, Dr. Kavitha Subramaniam

In Credit decisioning (Tier 1), Rebuilt from SAS

Strictly Confidential

Customer-level data under banking secrecy and PDPA, and models that decide on customers.

Named-user entitlement with the data owner's consent and a stated purpose; reviewed on expiry.

Tier 1 model: Material model: independent validation before use and annual revalidation.
Credit decision: Outputs feed a decision on a customer's credit; fairness and explainability tests apply.
Banking secrecy: Customer account information protected under the Financial Services Act 2013.
Customer PII: Contains or processes personal data about customers (PDPA 2010).

Try it

Live sandbox

application=PL-2026-33810

fixture

Evaluation only: restricted models run against synthetic or masked sample data until your business unit's access request is approved.

Feedback

Details

Updated: 2026-07-01
Latest version: 4.1.0
Licence: BU Restricted
Access: Restricted
Framework: scikit-learn
Language: N/A

Trained on

loan-applications-history →customer-360-mart →