Credit Underwriting

Credit underwriting is the process by which a lender evaluates a request for credit and decides whether to approve it, for how much, at what price and under what conditions. The underwriter gathers information about the borrower, analyses the risk that the money will not be repaid, and weighs that risk against the lender's policy and the return on offer. It applies to every form of lending, from credit cards and mortgages to corporate loans and trade credit.

Underwriting is the decision-making core of credit origination. The word is shared with other fields, and the general idea is described under underwriting: in insurance underwriting the risk assessed is a claim, and in securities markets an underwriter commits to buy an issue. In lending, the risk assessed is default.

A long-standing framework groups the analysis under five headings, often called the five Cs of credit.

• Character. The borrower's record of meeting obligations, drawn from the credit report, payment history and, for businesses, the reputation of owners and managers.
• Capacity. The ability to service the debt from income or cash flow, measured by ratios such as the debt-to-income ratio for individuals and the debt service coverage ratio for companies.
• Capital. The borrower's own resources and stake in the venture, including savings, equity and leverage.
• Collateral. Assets pledged as security and their value relative to the loan, expressed as the loan-to-value ratio.
• Conditions. The purpose of the loan, its terms, and the economic and industry environment in which it will be repaid.

High-volume consumer lending is underwritten mainly by automated decision systems. An application is checked against policy rules, such as minimum age or maximum indebtedness, and scored with a statistical model built on past repayment data. Applications above a cut-off credit score are approved, those below are declined, and borderline cases are referred to a person. Open banking data and transaction histories are increasingly used alongside bureau data, particularly for applicants with thin credit files.

Commercial and corporate lending relies more on judgement. An analyst reviews audited accounts, management forecasts, sector outlook and the structure of the proposed facility, tests how the borrower would cope under adverse scenarios, and assigns an internal rating that maps to a probability of default. The written analysis goes to an approver with authority under the credit approval matrix.

Underwriting produces more than a yes or no. The lender may approve a lower amount, shorten the term, require a guarantor or security, add covenants or raise the interest rate to reflect risk, a practice known as risk-based pricing. Approvals are often conditional on verification of income, a property valuation carried out under the lender's collateral valuation for lending standards, or other documents. A decline may have to be explained to the applicant under consumer credit law, and many lenders allow an appeal or a fresh review when new information is supplied. Each decision, with its reasons and any override of the score, is recorded so that it can be audited later.

Lenders set out their underwriting standards in a credit policy approved by the board, and supervisors review both the policy and how closely it is followed. Responsible lending rules in many countries require an assessment of affordability before consumer credit is granted, and mortgage rules often cap loan-to-value or loan-to-income ratios. Fair lending and anti-discrimination laws forbid decisions based on protected characteristics, which also constrains the variables that scoring models may use. Where models drive decisions, model risk rules call for independent validation and monitoring.

Accounting and capital rules depend on underwriting outputs. The ratings and scores assigned at approval feed the calculation of expected credit loss allowances and, for banks using internal models, regulatory capital.

Underwriting standards tend to loosen in good times, when competition is strong and recent losses are low, and to tighten sharply after losses appear. Typical failings include reliance on unverified income, optimistic collateral values, lending against rising asset prices rather than cash flow, and frequent overrides of policy. Models can also fail when conditions move outside the data they were built on. Lenders address these risks through independent credit review, limits on policy exceptions, monitoring of each cohort of new loans, and checks for fraudulent loan application patterns.