toolbox
Home / Legal, Safety & Compliance / AML Monitoring Audit Sample Size Estimator

AML Monitoring Audit Sample Size Estimator

The AML Monitoring Audit Sample Size Estimator estimates a statistical sample size for reviewing a finite population of aml monitoring records or cases. It uses a standard proportion-sampling formula with finite-population correction, allowing audit teams to choose population size, confidence level, expected exception rate, and margin of error. The output is a planning aid for audit design, not a regulator-mandated minimum. Risk-based compliance testing may require stratification, targeted selections, judgmental samples, or larger coverage than a purely statistical estimate.

Inputs

records
%
%
Result
—
calculated result
Initial sample (unlimited population)—
Finite-population sample—
Population sampled—
Expected exceptions in sample—

1. Define the population. Enter the number of records or cases in the audit population after applying the scope criteria.

2. Choose confidence. Select the confidence level used for the statistical planning scenario.

3. Set an expected exception rate. Use prior testing, control history, or a documented planning assumption rather than a fabricated benchmark.

4. Set the margin of error. A smaller margin increases the required sample because the estimate must be more precise.

5. Review and round up. The calculator rounds the finite-population result upward to a whole record so the planned sample is not understated.

For an expected exception proportion p, confidence z-score z, and margin of error e:

n₀ = z² × p × (1 − p) ÷ e²
n = n₀ ÷ (1 + (n₀ − 1) ÷ N)

Where N is the finite population size. Percent inputs are converted to decimals before calculation. The final sample is rounded up. This statistical model estimates precision for a proportion; it does not replace risk-based selections or requirements to test particular high-risk items.

What the result means

Use the result as a planning estimate based on the assumptions entered. Revisit the inputs when workload, legal scope, risk profile, staffing, or cost conditions change.

This tool provides general planning information and does not replace legal advice, a regulator-specific methodology, or an organization’s approved compliance procedures.

Given: 18,000 AML monitoring cases, 95% confidence, 4% expected exception rate, and 2.5% margin of error.

Calculation: n₀ = 1.96² × 0.04 × 0.96 ÷ 0.025² ≈ 236.03. Finite correction gives n ≈ 232.99, so round up to 233 records.

Result: A 233-record statistical sample fits these assumptions before adding any targeted high-risk selections.

Can this sample be used for transaction-monitoring model validation?

Not by itself. Model validation often requires broader methodology, data, tuning, scenario, and effectiveness testing. This calculator only estimates a proportion-based sample for a defined population.

Should suspicious or high-risk alerts be sampled randomly?

Not necessarily. High-risk alerts can be tested separately or at a higher rate while a random sample addresses the general population. Document the strata and selection logic so conclusions are traceable.

Why use an expected exception rate in AML audit planning?

It represents the proportion of the population expected to contain the condition being measured. Prior testing or a documented risk assumption is more defensible than inserting an industry average without evidence.

What happens when I choose a smaller margin of error?

The required statistical sample increases because the estimate is being asked to be more precise. The operational cost of that larger sample should be balanced against audit objectives and risk.

Does the sample-size result replace examiner judgment?

No. FFIEC materials emphasize risk-focused testing. Examiners and independent testers can expand, target, or otherwise adjust testing based on risk, findings, systems, and the quality of controls.