4.3 Audit Sampling: Statistical, Monetary Unit Sampling & Sampling Risk (ISA 530)

4.3 Audit Sampling: Statistical, Monetary Unit Sampling & Sampling Risk (ISA 530)

Core Principle: In auditing multi-billion euro institutional budgets, testing every transaction is neither possible nor cost-effective. Under ISA 530 and ISSAI 2530, audit sampling provides the mathematical and methodological foundation for drawing defensible inferences about an entire population from a representative sample. In public spending audits—such as the European Court of Auditors' Statement of Assurance (DAS)—sampling rigor is paramount to withstand institutional, parliamentary, and legal scrutiny.

---

The Definition and Scope of Audit Sampling

Under ISA 530 (Audit Sampling), audit sampling is defined as:

The application of audit procedures to less than 100% of items within a population of audit relevance such that all sampling units have a chance of selection in order to provide the auditor with a reasonable basis on which to draw conclusions about the entire population.

What is NOT Audit Sampling?

Auditors must carefully distinguish true audit sampling from other selective testing techniques:

---

Sampling Risk versus Non-Sampling Risk

The auditor faces two fundamental categories of risk when conducting testing procedures:

1. Sampling Risk

Sampling risk is the risk that the auditor's conclusion based on a sample may be different from the conclusion that would be reached if the entire population were subjected to the same audit procedure. Sampling risk leads to two distinct types of erroneous conclusions:

                                  THE SAMPLING RISK MATRIX

                                     Actual State of the Population
                              +--------------------------+--------------------------+
                              |    Population Correct    |   Population Materially  |
                              |   (Controls Effective)   |   Misstated / Deficient  |
    +-------------------------+--------------------------+--------------------------+
    | Auditor Decides:        |                          |     TYPE II ERROR        |
    | ACCEPT POPULATION /     |     CORRECT DECISION     |  (Beta Risk / Overreliance)
    | CONTROLS EFFECTIVE      |                          |  * Impacts EFFECTIVENESS *
    |                         |                          |  * Inappropriate Opinion *
    +-------------------------+--------------------------+--------------------------+
    | Auditor Decides:        |      TYPE I ERROR        |                          |
    | REJECT POPULATION /     | (Alpha Risk / Underrely) |     CORRECT DECISION     |
    | CONTROLS DEFICIENT      |  * Impacts EFFICIENCY *  |                          |
    |                         |  * Unnecessary Work *    |                          |
    +-------------------------+--------------------------+--------------------------+

The Critical Audit Distinction: Effectiveness versus Efficiency

• Type II Error (Beta Risk / Risk of Incorrect Acceptance / Overreliance): The auditor concludes that a population is materially correct when it is actually misstated, or that controls are effective when they are deficient. This directly affects AUDIT EFFECTIVENESS and is dangerous because it leads to an unjustified unmodified audit opinion. On competitive exams, this is recognized as the ultimate audit failure.
• Type I Error (Alpha Risk / Risk of Incorrect Rejection / Underreliance): The auditor concludes that a population is materially misstated when it is actually correct, or that controls are deficient when they work. This affects AUDIT EFFICIENCY because it prompts the auditor to perform unnecessary additional testing before ultimately arriving at the correct conclusion.

2. Non-Sampling Risk

Non-sampling risk is the risk that the auditor reaches an erroneous conclusion for any reason not related to sampling risk. Causes include: applying inappropriate audit procedures; misinterpreting documentary evidence; failing to recognize an obvious fraudulent anomaly; or relying on erroneous management representations. Non-sampling risk cannot be mathematically quantified; it is controlled through rigorous audit planning, supervision, and quality management systems (ISA 220 / ISQM 1).

---

Statistical versus Non-Statistical Sampling

ISA 530 recognizes two overarching sampling frameworks, both of which require professional competence:

1. Statistical Sampling: An approach to sampling that has the following characteristics:
   • Random selection of the sample items; and
   • The use of probability theory to evaluate sample results, including measurement of sampling risk.
2. Non-Statistical (Judgmental) Sampling: Any sampling approach that does not possess both characteristics of statistical sampling. The auditor uses professional judgment to determine sample size, select items, and evaluate results.

Important Principle: While statistical sampling quantifies sampling risk mathematically, non-statistical sampling does not. However, both approaches require the auditor to design a representative sample and project identified misstatements to the population.

---

Sample Selection Techniques

Auditors deploy several techniques to select sampling units from a population:

• Random Selection: Applied using random number generators where each sampling unit in the population has an equal mathematical probability of selection.
• Systematic Selection: The number of sampling units in the population is divided by the sample size to give a sampling interval ($k = N / n$). Having established a random start within the first interval, every $k$-th item thereafter is selected.
• Stratified Selection: Subdividing a heterogeneous population into homogeneous subpopulations (strata) based on specific characteristics (such as monetary value or risk profile). Each stratum is sampled separately, allowing the auditor to direct more testing effort to high-value or high-risk items while reducing total sample size.
• Haphazard Selection: The auditor selects the sample without following a structured mathematical technique, avoiding conscious bias or predictability. Haphazard selection is strictly non-statistical because sampling probabilities cannot be computed.
• Block Selection: Selecting contiguous items from within a population (e.g., all vouchers issued in the third week of November). Block selection is rarely appropriate in financial auditing because transactions in close sequence typically share similar characteristics, creating high non-representativeness.

---

Monetary Unit Sampling (MUS) in EU Public Spending Audits

Monetary Unit Sampling (MUS)—also referred to as Probability Proportional to Size (PPS) or Dollar-Unit Sampling—is a statistical sampling method that may be used in financial and public-spending audits, including by the European Court of Auditors (ECA) where appropriate when establishing the annual Statement of Assurance (DAS) on the legality and regularity of EU budgetary expenditure.

The Operating Principles of MUS

1. Individual Euro as Sampling Unit: Instead of treating physical vouchers or project files as sampling units, MUS defines every individual euro in the population as an independent sampling unit.
2. Probability Proportional to Size: A payment transaction of EUR 1,000,000 contains 1,000,000 sampling units, whereas a transaction of EUR 10,000 contains only 10,000. Therefore, the EUR 1,000,000 transaction is 100 times more likely to be selected for audit.
3. High-value inclusion: In a systematic monetary-unit design, items at or above the interval will ordinarily have certainty of inclusion and may contain multiple selected monetary units. The auditor still documents how high-value items are treated and how results are evaluated under the chosen design.

The Mechanics of MUS Selection

To execute MUS systematically:

1. Calculate the Sampling Interval ($SI$): $SI = \frac{\text{Total Population Book Value (EUR)}}{n \text{ (Sample Size)}}$
2. Generate Cumulative Monies: Arrange the population and compute a running cumulative monetary total.
3. Select Items: Establish a random start between 1 and $SI$. Subsequent sampling units are selected at fixed steps of $+SI$. The physical transaction containing the selected euro is audited.

---

Evaluating Sample Results

For a test of details, ISA 530 requires the auditor to project misstatements found in the sample to the population. The projection method must fit the sampling design. For a test of controls, the auditor evaluates the observed deviation rate and sampling risk rather than projecting a monetary misstatement.

In monetary-unit sampling, a misstatement can be expressed as a tainting percentage:

$\text{Tainting} = \frac{\text{Misstatement}}{\text{Recorded amount}}$

If a sampled EUR 10,000 item contains EUR 2,500 of unsupported cost, the tainting is 25%. If the applicable sampling interval is EUR 50,000 and the item is below that interval, a simplified practice projection is EUR 12,500. Actual evaluation depends on the design, high-value items, confidence level and the auditor's documented methodology.

The auditor then considers the projected result together with the allowance for sampling risk, qualitative features of the errors, anomalies, population definition and evidence from other procedures. An estimate is not automatically a fraud amount or recoverable financial loss. For ECA regularity work, the 2% quantitative materiality threshold is a key benchmark, but the opinion also reflects confidence bounds and the nature and pervasiveness of error.