An auditor selects files for a delegated authority review by combining risk-based targeting with a smaller element of random sampling. The auditor first reviews the bordereaux, claims listings and complaints data supplied by the coverholder or MGA, then picks files that reflect where risk is concentrated. The aim is a sample that is representative, defensible and focused on real exposure.
How that sample is built depends on the type of authority being audited, the size of the portfolio and the insurer's own audit scope. The sections below explain the main questions underwriting, compliance and delegated authority managers ask about file selection.

Is file selection random or risk-based in a delegated authority audit?

File selection in a delegated authority audit is primarily risk-based, usually supported by a smaller random element. The auditor targets files where errors would be most costly or most likely, then adds random files to test the wider population fairly.
A purely random sample can miss the very files that matter most. A single large, unusual or late-reported claim tells an auditor far more than ten routine policies. Equally, a sample made up only of high-risk files can give a distorted picture of everyday performance.
A blended approach solves both problems:

• Targeted selection focuses on files that carry higher financial, regulatory or conduct risk.
• Random selection tests the underlying population and shows whether good practice is consistent.

This also keeps the review proportionate. The auditor is not trying to test everything, only to form a reliable view of how the delegated authority operates.

What data do auditors request before choosing files?

Before choosing files, auditors request complete bordereaux, claims listings, complaints registers and details of the underlying binding authority. This data lets the auditor see the full population of risks, claims and complaints, so the sample is drawn from a reliable source rather than one the coverholder has pre-filtered.
Typical requests include:

1. Premium and risk bordereaux covering the audit period.
2. Claims bordereaux and a list of open, closed and declined claims.
3. The complaints log, including any complaints referred to the insurer or an ombudsman.
4. The binding authority agreement, underwriting guidelines and referral limits.
5. Details of any cancellations, endorsements and premium adjustments.

Auditors will usually compare this information against what the insurer holds. Gaps or mismatches between the two datasets are themselves useful signals, and they often shape which files are pulled for closer testing.

Which files are most likely to be selected for testing?

The files most likely to be selected are those that carry higher risk or sit close to the limits of the authority. This includes large premium risks, high-value or contentious claims, referrals, declined claims and complaints, along with files that show unusual patterns in the data.
Common selection triggers in delegated underwriting include risks near or above authority limits, unusual classes or territories, and policies with significant manual adjustments to rating. In a delegated claims audit, auditors often look at:

• Large or complex claims and those close to the settlement authority limit.
• Declined or repudiated claims, where fair treatment is a key concern.
• Claims with long delays between notification and first action.
• Claims with reserve movements that look inconsistent.

Complaints files are almost always reviewed, because they show how the firm responds when something has gone wrong.

How many files does an auditor typically review?

There is no single fixed number. The sample size depends on the size and complexity of the portfolio, the insurer's audit requirements and the level of risk identified. As a general principle, larger or riskier books receive larger samples, while smaller, stable portfolios need fewer files.
Auditors usually agree the scope with the insurer in advance. Factors that influence sample size include:

• The volume of policies and claims in the audit period.
• The number of products, classes and territories covered.
• Previous audit results and any outstanding actions.
• The time available on site or through remote review.

A sample is not meant to prove that every file is correct. It gives the insurer reasonable assurance about how the delegated authority is run, and it should be large enough to show whether any issue is isolated or systemic.

What happens when file testing uncovers a problem?

When file testing uncovers a problem, the auditor widens the review to establish whether it is a one-off error or a recurring weakness. Additional files are tested, the root cause is identified, and the finding is reported with a clear rating and recommended corrective action.
The process generally follows a logical path. First, the auditor confirms the facts with the coverholder or MGA, so there is no misunderstanding. Next, further files of a similar type are tested to gauge how widespread the problem is. The finding is then explained in the report, including why it matters and what the firm should do about it.
A well-written report distinguishes between isolated slips and control failures. That distinction matters, because it determines whether the insurer needs a simple correction or a formal action plan with agreed deadlines.

Can a coverholder or MGA influence which files are chosen?

A coverholder or MGA should not influence which files are chosen, because that would undermine the independence of the audit. The auditor controls selection. The firm's role is to supply complete and accurate data, and to retrieve the selected files promptly once they are requested.
Auditors usually make the final selection from the full dataset, often shortly before or during the review, so the firm cannot prepare only certain files. Transparency helps everyone. Firms that maintain well-organised records, clear file notes and consistent referral evidence tend to have smoother audits, because the auditor can follow decisions without lengthy queries.
If a firm believes a selected file is unrepresentative, the right approach is to explain the context to the auditor, not to ask for a replacement.

How Otter Audit helps with delegated authority file selection

Otter Audit carries out delegated authority audits for insurers across the Lloyd's, London and company markets, with ACII-qualified auditors who have reviewed delegated underwriting, claims, pre-bind, run-off and complaints files in many territories. Our approach to file selection is built to give insurers dependable assurance.
Our audits offer:

• A documented, risk-based sampling approach agreed with the insurer.
• Clear, actionable reports that explain each finding and the reasoning behind it.
• Reports delivered on time, supported by a six-year record of meeting agreed deadlines.
• Technical knowledge across a wide range of specialist and non-standard products.

If you would like to discuss the scope or sampling approach for your next review, please contact our team.

Related Articles

• What information is needed before a coverholder audit begins?
• Can an insurance audit be completed entirely remotely?
• How do auditors conduct interviews with coverholder staff?
• How do auditors assess compliance with a binding authority agreement?
• How can insurers tell whether an audit methodology is sufficiently thorough?