Underwriting

The purpose of underwriting

Underwriting is the process of assessing and pricing the risk of a lending decision, and deciding whether to accept it and on what terms. It is where a lender brings together everything it knows about an applicant, their credit history, income, affordability, the purpose of the loan and any security, and reaches a judgement: lend or decline, how much, over what term, at what rate and with what conditions. The word carries both halves of that judgement, assessing the risk and pricing it.

Underwriting is distinct from the earlier marketing and application steps and from the later collections activity. It is the decision point, and the quality of underwriting largely determines how a book will perform. A disciplined underwriting process is the single most important defence a lender has against future credit losses.

Automated and manual underwriting

Most high-volume consumer lending is underwritten automatically. A decisioning engine applies scorecards and policy rules to bureau and application data and, increasingly, to Open Banking affordability information, returning a decision in seconds. Automated underwriting is consistent, fast and auditable, and it handles the bulk of clean-cut cases well.

Manual underwriting still has its place, particularly for larger, more complex or marginal cases, and in much of SME and specialist lending. A human underwriter can weigh context that a model handles poorly, an unusual income pattern, a one-off historical blip, the commercial story behind a business loan, and exercise judgement within policy. Many lenders use a hybrid model, auto-deciding where they can and referring edge cases to an underwriter. The art is in setting the referral thresholds: too tight and underwriters are swamped with cases the engine could have handled, too loose and genuinely marginal decisions slip through without the human judgement they needed.

• Automated: scorecards and rules on bureau, application and Open Banking data for speed and consistency
• Manual: human judgement for complex, large or borderline cases
• Referral: automated screening that routes unclear cases to an underwriter

What underwriting weighs

A sound underwriting decision rests on more than a credit score. The underwriter, human or automated, considers the applicant's repayment history and conduct, their affordability once existing commitments and essential costs are accounted for, their resilience to a change in circumstances, and, where relevant, the value and quality of any security. Identity and KYC checks confirm the lender is dealing with who it thinks it is and that the lending is lawful.

Affordability has become central to UK underwriting, driven by the FCA's expectations under CONC and the Consumer Duty. It is no longer enough to establish that an applicant is creditworthy in the sense of likely to repay; the lender must also be satisfied the borrowing is affordable and sustainable for that customer. Open Banking data has sharpened this considerably by replacing declared figures with evidenced income and spending.

Underwriting, consistency and fairness

Because underwriting decides who gets credit and on what terms, it must be consistent, explainable and fair. Lenders are expected to be able to explain a decline, to avoid unlawful discrimination, and to test their models and policies for unintended bias. Clear policy, good record-keeping and regular review of outcomes are what keep underwriting defensible.

Underwriting is also where good outcomes are designed in. Declining an unaffordable loan, or offering a smaller amount or longer term that the customer can sustain, is good underwriting under the Consumer Duty as much as it is good risk management. The better the data and the clearer the policy, the more confidently a lender can say yes to the right applicants and no to the ones who would come to harm.