The evidence-based process of deciding whether and on what terms to extend, renew, or modify credit under applicable policy and law.
Credit underwriting is the evidence-based process of deciding whether and on what terms to extend, renew, or modify credit. It evaluates the borrower, repayment source, amount, payment structure, collateral, guarantees, conditions, and downside risk under applicable policy and law.
Underwriting can be automated, manual, or a combination. A credit score or model output may inform the decision, but it does not replace accurate inputs, policy controls, approval authority, documentation, or legally supportable reasons.
Identify the applicant, amount, purpose, product, term, payment schedule, rate structure, collateral, guarantees, and requested timing. Include existing and contingent exposure to related parties where policy requires aggregation.
Evidence may include identity, income, employment, tax returns, financial statements, bank records, credit reports, debt schedules, receivables aging, appraisals, contracts, and ownership documents. Requirements vary by product and jurisdiction.
For a consumer, repayment may come from recurring employment or other verified income. For a business, it may come from operating cash flow, asset conversion, contract receipts, or project cash flow. The source should match the timing and currency of the obligation.
The underwriter evaluates recurring cash available after obligations, prior payment performance, leverage, liquidity, volatility, and other relevant evidence. Historical performance should be adjusted for one-time, unsupported, or noncash items where appropriate.
Assess whether amount, maturity, amortization, covenants, collateral, guarantees, and conditions address the identified risks. A tighter structure does not make an unaffordable loan affordable.
Stress the assumptions most likely to weaken repayment, such as lower sales, income interruption, rate reset, customer loss, collateral decline, or higher operating costs.
Identify mandatory rules, guidelines, exceptions, model overrides, concentration limits, and required approvers. The decision maker should receive the complete risk picture.
Record the evidence, calculations, assumptions, decision, conditions, exceptions, and actual reasons. Complete applicable notices and retain records as required.
| Area | Consumer credit | Commercial credit |
|---|---|---|
| Applicant evidence | Income, employment, assets, debts, credit history | Financial statements, tax returns, debt schedule, ownership, projections |
| Repayment analysis | Payment relative to verified income and obligations | Operating cash flow, liquidity, leverage, debt service, business risks |
| Qualitative factors | Stability and relevant credit behavior | Management, industry, competition, customers, suppliers, strategy |
| Structure | Amount, term, rate, payment, security | Facilities, maturity, amortization, borrowing base, covenants, guarantees |
| Monitoring | Payment performance and account behavior | Financial reporting, covenants, collateral, borrowing base, risk grades |
| Legal and policy overlay | Consumer disclosure and fair-lending rules may apply | Entity, guaranty, collateral, covenant, and commercial-law issues may dominate |
The categories can overlap. A small-business request may depend on both business cash flow and personal guarantees.
Assume a business requests a $250,000 equipment loan. The proposed annual principal and interest payment is $62,500. Normalized annual cash flow available for debt service (CFADS) is estimated at $108,000.
The simplified base debt service coverage ratio (DSCR) is:
Base DSCR = $108,000 / $62,500 = 1.73x
If a downside case reduces CFADS by 25%:
Stressed CFADS = $108,000 x 75% = $81,000
Stressed DSCR = $81,000 / $62,500 = 1.30x
Assume the equipment’s estimated orderly liquidation value is only $150,000. The collateral would not cover the original principal if default occurred immediately, before collection costs and other claims.
A defensible decision should therefore examine:
The ratios do not dictate approval. They show that cash flow, stress resilience, and collateral recovery answer different questions.
| Method | Strength | Main control need |
|---|---|---|
| Automated rules | Fast, repeatable application of explicit criteria | Accurate data, tested rules, change control, exception logging |
| Statistical model | Consistent risk estimation across many applications | Validation, monitoring, data relevance, limits, governance |
| Manual review | Can evaluate complex evidence and unusual structures | Training, consistency, documentation, conflicts, second review |
| Hybrid process | Combines scale with human review of referrals or exceptions | Clear handoffs, override standards, reason tracking |
Current interagency model-risk guidance emphasizes risk-based governance, validation, monitoring, effective challenge, documentation, and oversight appropriate to model use and materiality. It does not make every spreadsheet or deterministic rule a statistical model.
An underwriter should know whether an automated result is an approval, a recommendation, a risk estimate, or a referral. Those outputs are not interchangeable.
Possible outcomes include:
For covered applications, Regulation B governs evaluation and notification requirements. Adverse-action reasons should be specific enough to identify the actual principal reasons and should match the factors used in the decision.
Underwriting is prospective and cannot guarantee repayment. Borrower information can be incomplete or false, models can deteriorate, forecasts can fail, collateral can lose value, and economic conditions can change rapidly. Controls reduce uncertainty but do not eliminate credit loss.
This page is educational and is not personalized lending, mortgage, investment, legal, compliance, accounting, or financial advice.