Credit Scoring Engine

The Credit Scoring Engine provides a configurable, weighted scorecard system for evaluating borrower creditworthiness during loan application processing. Scores are calculated automatically from client data, with manual overrides available.

Navigation: Administration → Products → Loan Products → Credit Score Cards

Workspace Reference

Workspace	How Users Reach It	What Users Usually Do There
Credit Score Cards	Loans -> Setup -> Products, then open credit score card setup where available	Create scorecard templates, criteria, ranges, and risk grades.
Loan Products	Loans -> Setup -> Products	Link or review the product context that decides when scoring is relevant.
Loan Applications	Loans -> Operations -> Applications	Run scoring during credit assessment and review the score result before approval.
Parties And Deposits	Parties and Deposits service workspaces	Maintain source data that auto-resolved criteria may use, such as age, tenure, savings balance, or existing obligations.

Overview

The engine follows the industry-standard 5C model (Character, Capacity, Capital, Conditions, Collateral) and supports:

• Weighted criteria — each factor carries a configurable weight
• Point ranges — map raw values to point scores (numeric ranges, categories, booleans)
• Risk grades — translate final scores into approval decisions (A–E)
• Auto-resolve — pull client data automatically from the system (age, tenure, DTI, balances)
• Manual input — officers can enter values for criteria that require judgement

┌──────────────┐     ┌──────────────┐     ┌──────────────┐
│  Score Card  │────▶│   Criteria   │────▶│    Ranges    │
│  (template)  │     │  (factors)   │     │ (value→pts)  │
└──────────────┘     └──────────────┘     └──────────────┘
       │
       ▼
┌──────────────┐     ┌──────────────┐
│    Grades    │────▶│   Decision   │
│  (A/B/C/D/E) │     │  (approve/   │
│              │     │   review/    │
│              │     │   reject)    │
└──────────────┘     └──────────────┘

Key Concepts

Concept	Description
Score Card	A named scorecard template with a score range (e.g. 0–1000) and version
Criterion	A single scoring factor (e.g. Client Age, DTI Ratio) with a weight and max points
Range	A mapping rule within a criterion: "if value is 26–35, award 70 points"
Grade	A risk tier defined by score thresholds: "600–799 = Grade B = Auto Approve"
Data Source	Where the raw value comes from — AUTO (system-resolved) or MANUAL (officer input)
Evaluation Type	How values are matched: NUMERIC_RANGE, CATEGORY, or BOOLEAN

Score Cards

Score cards are managed from the Credit Score Cards tab on the Loan Products page.

Score Card Fields

Field	Description	Required
Card Name	Descriptive name (e.g. "Standard Risk Card")	Yes
Model Version	Version tag (e.g. "v1.0")	Yes
Score Min	Minimum possible score (default 0)	No
Score Max	Maximum possible score (default 1000)	No

Actions

• Create — add a new score card
• Edit — modify card name/version/range
• Configure — open the criteria & grades management dialog
• Delete — remove a score card

Configuring a Score Card

Click the Configure (⚙) action on a score card row to open the configuration dialog with two tabs:

Scoring Criteria Tab

Each criterion represents a factor in the credit assessment.

Field	Description
Code	Unique identifier (e.g. CLIENT_AGE, DTI_RATIO)
Name	Display name
Category (5C)	CHARACTER, CAPACITY, CAPITAL, CONDITIONS, COLLATERAL, or CUSTOM
Evaluation Type	NUMERIC_RANGE, CATEGORY, or BOOLEAN
Data Source	AUTO (system resolves) or MANUAL (officer enters)
Auto Source	When AUTO, which system data point to pull (see Data Sources below)
Weight	Relative importance (0.00–1.00). All weights should sum to 1.0
Max Points	Maximum points this criterion can award
Default Points	Points when no range matches (fallback)
Required	Whether this criterion must have a value to evaluate

Point Ranges

Each criterion has one or more ranges that map raw values to point scores:

For NUMERIC_RANGE:

Range Label	Min (≥)	Max (<)	Points
18–25	18	25	30
26–35	26	35	70
36–50	36	50	100
51+	51	120	60

For CATEGORY: Each range specifies a category value (e.g. "MALE", "MARRIED").

For BOOLEAN: Each range specifies true (1) or false (0).

Risk Grades Tab

Grades map final scores to risk tiers and approval decisions.

Field	Description
Code	Grade code (e.g. A, B, C, D, E)
Name	Display name (e.g. "Excellent", "Good")
Min Score	Lower bound of score range (inclusive)
Max Score	Upper bound of score range (inclusive)
Decision	AUTO_APPROVE, MANUAL_REVIEW, or AUTO_REJECT
Rate Adj (bps)	Interest rate adjustment in basis points (e.g. +50bps for Grade B)
Max LTV %	Maximum loan-to-value percentage (optional)
Color	Hex color for UI display

Example Grade Scale (0–1000):

Grade	Range	Decision	Rate Adjustment
A — Excellent	800–1000	Auto Approve	+0 bps
B — Good	600–799	Auto Approve	+50 bps
C — Fair	400–599	Manual Review	+150 bps
D — Poor	200–399	Manual Review	+300 bps
E — Very Poor	0–199	Auto Reject	+500 bps

Available Data Sources

When a criterion uses AUTO data source, the engine resolves the value from client/application data at evaluation time. Available sources:

Code	Label	Category	Type
CLIENT_AGE	Client Age (years)	CHARACTER	NUMERIC_RANGE
CLIENT_GENDER	Client Gender	CHARACTER	CATEGORY
MARITAL_STATUS	Marital Status	CHARACTER	CATEGORY
CUSTOMER_TENURE_MONTHS	Customer Tenure (months)	CHARACTER	NUMERIC_RANGE
EXISTING_LOAN_COUNT	Active Loan Count	CHARACTER	NUMERIC_RANGE
ARREARS_LOAN_COUNT	Loans in Arrears Count	CHARACTER	NUMERIC_RANGE
DTI_RATIO	Debt-to-Income Ratio	CAPACITY	NUMERIC_RANGE
AFFORDABILITY_RATIO	Affordability Ratio	CAPACITY	NUMERIC_RANGE
MONTHLY_GROSS_INCOME	Monthly Gross Income	CAPACITY	NUMERIC_RANGE
MONTHLY_NET_INCOME	Monthly Net Income	CAPACITY	NUMERIC_RANGE
MONTHLY_EXPENSES	Monthly Expenses	CAPACITY	NUMERIC_RANGE
DISPOSABLE_INCOME	Disposable Income	CAPACITY	NUMERIC_RANGE
SAVINGS_BALANCE	Savings / Deposit Balance	CAPITAL	NUMERIC_RANGE
TOTAL_OUTSTANDING_DEBT	Total Outstanding Debt	CAPITAL	NUMERIC_RANGE
COLLATERAL_COVERAGE	Collateral Coverage Ratio	COLLATERAL	NUMERIC_RANGE
REQUESTED_AMOUNT	Requested Loan Amount	CONDITIONS	NUMERIC_RANGE

Evaluating a Loan Application

During loan application processing, the Credit Scoring stage includes the scoring engine:

Workflow

1. Select Score Card — choose which scorecard to evaluate against
2. Click Evaluate — the engine:
   • Loads all active criteria for the selected card
   • For each AUTO criterion, resolves the value from client/application data
   • Matches each value against configured ranges to determine points
   • Calculates weighted score: Σ(points × weight) / Σ(max_points × weight) × score_max
   • Looks up the matching grade based on the final score
3. Review Breakdown — the evaluator displays:
   • Per-criterion results (raw value, matched range, points, weight, weighted score)
   • Total score, weighted score, assigned grade, and decision
4. Auto-Fill — results automatically populate the form fields:
   • Score, Risk Grade, Pass/Fail, Application Score
   • Credit Assessment JSON (full breakdown for audit)
5. Override — officers can manually adjust any field before submitting

Score Calculation

For each criterion:
  weighted_points = points_earned × weight

Total weighted = Σ weighted_points
Maximum weighted = Σ (max_points × weight)

Final score = (Total weighted / Maximum weighted) × score_max

Example:

Criterion	Points	Weight	Weighted
Client Age (32 yrs → "26–35")	70	0.30	21.0
DTI Ratio (0.28 → "Good 20–35%")	75	0.40	30.0
Customer Tenure (18 mo → "1–3 years")	80	0.30	24.0
Totals		1.00	75.0 / 100.0

Final score = 75.0 / 100.0 × 1000 = 750 → Grade B (Good) → Auto Approve

Permissions

Permission	Description
CREATE_CREDIT_SCORE_CARD	Create new credit score card templates
UPDATE_CREDIT_SCORE_CARD	Edit score cards, manage scoring criteria, ranges, and risk grades
DELETE_CREDIT_SCORE_CARD	Remove score cards and their associated criteria and grades
VIEW_CREDIT_SCORE_CARD	View score cards, criteria, ranges, grades, and data sources
EVALUATE_CREDIT_SCORECARD	Run the scoring engine during loan application processing
VIEW_CREDIT_SCORE_HISTORY	View client credit score history records

Operational Controls

When configuring score cards, administrators should confirm:

• every criterion has a clear business meaning
• required criteria are only used where the value is always available or must be entered manually
• score ranges do not overlap in confusing ways
• grades cover the full score range
• auto-approve and auto-reject grades match credit policy
• manual overrides are reviewed through the normal approval process

Credit scoring is decision support. It should not replace credit officer judgement where policy requires manual review.

Best Practices

1. Weights should sum to 1.0 — while not strictly enforced, the weighted calculation is most meaningful when total weight equals 1.0
2. Cover the full score range with grades — ensure no gaps between grade min/max boundaries
3. Start with 3–5 criteria — the 5C model provides a natural grouping; avoid over-engineering with too many factors
4. Use AUTO sources where possible — reduces manual data entry and ensures consistency
5. Version your score cards — create new versions (v1.1, v2.0) rather than modifying active cards, so historical scores remain traceable
6. Review grade thresholds periodically — calibrate against actual loan performance data
7. Set Default Points conservatively — when no range matches (data unavailable), a low default prevents inflated scores

Troubleshooting

Issue	Cause	Fix
Evaluate button disabled	No score card selected	Select a score card from the lookup
Score returns 0	All criteria have no matching ranges	Check that ranges cover the expected value space
AUTO source returns NULL	Client data missing (e.g. no date_of_birth)	Ensure client profile is complete, or use MANUAL
Grade shows "—"	Final score falls outside all grade ranges	Ensure grades cover the full score_min to score_max
Weights don't sum to 1.0	Criteria weights misconfigured	Edit criteria and adjust weights

Related Topics

• Loan Products — product configuration
• Loan Operations — disbursement, repayment, write-off
• Loan Accounts — account lifecycle