7.4 Substantive Analytical Procedures (AU-C 520)

7.4 Substantive Analytical Procedures (AU-C 520)

Core Principle: Under AU-C 520 (Analytical Procedures), analytical procedures consist of evaluations of financial information made by a study of plausible relationships among both financial and non-financial data. Analytical procedures also encompass the investigation of identified fluctuations or relationships that are inconsistent with other relevant information or that differ from expected values by a significant amount. A basic premise underlying analytical procedures is that plausible relationships among data may reasonably be expected to exist and continue in the absence of known conditions to the contrary.

---

1. The Three Audit Stages of Analytical Procedures

One of the most frequently tested concepts on the CPA AUD exam is distinguishing between the stages where analytical procedures are mandatory versus where they are discretionary (optional).

                         THE THREE STAGES OF ANALYTICAL PROCEDURES

      STAGE 1: PLANNING                  STAGE 2: SUBSTANTIVE                STAGE 3: FINAL REVIEW
   (Risk Assessment - AU-C 315)         (Fieldwork - AU-C 330/520)         (Overall Conclusion - AU-C 520)
  +----------------------------+      +----------------------------+      +----------------------------+
  |        MANDATORY           |      |  DISCRETIONARY / OPTIONAL  |      |        MANDATORY           |
  |                            |      |                            |      |                            |
  | - Purpose: Identify areas  |      | - Purpose: Detect material |      | - Purpose: Form overall    |
  |   of heightened RMM and    |      |   misstatements at the     |      |   conclusion whether F/S   |
  |   unusual transactions     |      |   assertion level          |      |   are consistent with      |
  | - Data: Highly aggregated, |      | - Data: Highly disaggre-   |      |   auditor's understanding  |
  |   high-level financial &   |      |   gated, detailed monthly  |      | - Data: Broad, high-level  |
  |   operating figures        |      |   data, non-financial data |      |   financial statement level|
  +----------------------------+      +----------------------------+      +----------------------------+

Detailed Comparison Table

Exam Trap (The Universal Requirement Trap): Substantive analytical procedures are never mandatory. The auditor may choose to test an account balance entirely through tests of details (e.g., physical inspection, confirmation, recalculation) without executing any substantive analytical procedures. Only risk assessment analytics and final overall-review analytics are required by GAAS.

---

2. The 4-Step Methodology of Substantive Analytical Procedures

When the auditor elects to use substantive analytical procedures, AU-C 520 dictates a rigorous four-step process:

                SUBSTANTIVE ANALYTICAL PROCEDURES: 4-STEP FRAMEWORK

    [ STEP 1: DETERMINE SUITABILITY ]
    - Assess assertion predictability
    - Balance efficiency vs. tests of details
                   |
                   v
    [ STEP 2: EVALUATE DATA RELIABILITY ]
    - Assess source independence
    - Test controls over underlying data (IPE)
                   |
                   v
    [ STEP 3: DEVELOP INDEPENDENT EXPECTATION ]
    - Develop expectation BEFORE viewing recorded balance
    - Maximize precision (disaggregation & non-financial drivers)
                   |
                   v
    [ STEP 4: COMPARE & INVESTIGATE DIFFERENCES ]
    - Compare recorded amount to expectation
    - If difference > Tolerable Threshold:
      * Inquire of management
      * Corroborate with audit evidence (Inquiry ALONE is never enough!)
      * Perform tests of details if uncorroborated

Step 1: Determine the Suitability of the Procedure

Substantive analytical procedures are not equally effective for all financial statement accounts or assertions:

• High Suitability: Large volumes of routine, predictable transactions that recur over time (e.g., rental revenue, straight-line depreciation, interest expense on fixed debt, utility expenses, standardized payroll).
• Low Suitability: Non-routine transactions, unique year-end transactions, accounts subject to substantial management discretion or subjective estimates (e.g., warranty liabilities, goodwill impairment, litigation reserves, Level 3 derivative fair valuations). These require direct tests of details.

Step 2: Evaluate the Reliability of the Data Used

The persuasiveness of the expectation depends entirely on the reliability of the underlying data:

1. Source Independence: Data obtained from independent sources outside the entity (e.g., regulatory filings, published market interest rates, third-party leasing indices) is far more reliable than internal data.
2. Controls Over Data Preparation: If internal data is used (for example, production units, occupancy counts, or sales volumes), the auditor considers the controls over its preparation and may test those controls or test the data directly for accuracy and completeness (information produced by the entity, or IPE).
3. Comparability: Industry data must be drawn from genuinely comparable peers, not broad, unstratified global averages.
4. Audit Status: Prior-year data that was audited provides greater reliability than unaudited historical numbers.

Step 3: Develop an Independent Expectation with Sufficient Precision

The Cardinal Rule of Substantive Analytics: The auditor MUST formulate an independent expectation BEFORE comparing it to the client's recorded balance. Taking the client's recorded general ledger amount and seeking to rationalize it after the fact is a critical auditing failure.

Factors Governing Expectation Precision

Precision is the measure of the closeness of the auditor's expectation to the true underlying amount. Greater precision produces more persuasive audit evidence:

1. Predictability of the Relationship:
   • Income statement accounts are more predictable than balance sheet accounts because the income statement represents transactions accumulated over a period of time, whereas the balance sheet represents a snapshot at a single point in time subject to timing anomalies.
   • Contractual relationships are more predictable than discretionary expenses. Interest expense (governed by loan agreements) and rent expense (governed by lease terms) are highly predictable. Advertising expense, travel & entertainment, and research & development are subject to arbitrary management discretion and are inherently difficult to predict precisely.
   • Transactions in a stable economic environment are more predictable than those in volatile or rapidly shifting industries.
2. Level of Disaggregation:
   • The more disaggregated the data, the higher the precision.
   • Example: Developing an expectation for sales monthly by individual store is vastly more precise than developing a single annual expectation for total nationwide sales.
   • The Danger of Aggregation (Offsetting Errors): Highly aggregated data can mask significant material misstatements through offsetting errors (e.g., Store A overstates sales by $2,000,000 and Store B understates sales by $2,000,000; an aggregated annual analysis reveals zero variance!).
3. Availability and Use of Non-Financial Data:
   • Incorporating reliable physical operational drivers (e.g., number of occupied rooms in a hotel, number of credit hours billed by a university, total passenger miles flown by an airline) paired with contract pricing yields the most precise expectations.

Step 4: Compare Recorded Amount to Expectation & Investigate Variances

A. Setting the Tolerable Difference Threshold

Before performing the comparison, the auditor must establish a threshold of difference that can be accepted without further investigation (the tolerable difference). This threshold:

• Is determined based on performance materiality and the desired level of assurance.
• Must be small enough that a material misstatement could not hide inside it, so it is set with performance materiality and tolerable misstatement in mind.

B. Investigating Significant Differences

If the difference between the recorded amount and the auditor's independent expectation exceeds the tolerable threshold, AU-C 520 mandates a two-step investigation:

1. Inquire of Management: Seek management's business explanation for the variance.
2. Obtain Corroborating Audit Evidence: Inquiry alone is NEVER sufficient. The auditor must independently substantiate management's verbal statements by inspecting supporting documentation (e.g., examining amended contracts, checking price increase authorizations, reviewing bank statements).
3. Failure to Corroborate: If management cannot provide a valid explanation, or if their explanation is uncorroborated by evidence, the auditor must perform expanded substantive tests of details to quantify the misstatement.

---

3. Four Types of Analytical Models

                                HIERARCHY OF ANALYTICAL PRECISION

      MODEL TYPE                METHODOLOGY                             PRECISION LEVEL
  +-------------------+-----------------------------------------------+-----------------+
  | 1. TREND ANALYSIS | Comparing account balances across prior years | LOW             |
  +-------------------+-----------------------------------------------+-----------------+
  | 2. RATIO ANALYSIS | Financial ratios compared to budget/industry  | MODERATE        |
  +-------------------+-----------------------------------------------+-----------------+
  | 3. REASONABLENESS | Mathematical models using operating drivers   | HIGH            |
  +-------------------+-----------------------------------------------+-----------------+
  | 4. REGRESSION     | Statistical modeling measuring variables      | HIGHEST         |
  +-------------------+-----------------------------------------------+-----------------+

1. Trend Analysis: Comparing a recorded balance over multiple historical periods (e.g., comparing 2026 office supply expense to 2025 and 2024). Provides low precision because it ignores changes in operating conditions.
2. Ratio Analysis: Computing relationships between financial statement line items (e.g., gross margin percentage, inventory turnover, current ratio) and comparing them to prior years or industry benchmarks. Provides moderate precision.
3. Reasonableness Testing (Predictive Modeling): Developing a formal mathematical expectation using operational relationships and rate drivers (e.g., calculating expected apartment rental income: Units × Occupancy Rate × Average Monthly Rent × 12 months). Highly persuasive.
4. Regression Analysis: Sophisticated statistical modeling that quantifies the exact mathematical relationship between independent economic/operational variables and dependent financial metrics. Provides the highest precision and statistical rigor.

---

4. Mandatory AU-C 520 Workpaper Documentation Rules

When an auditor performs substantive analytical procedures, AU-C 520 requires documentation of three items: the expectation and the factors considered in developing it (when not otherwise readily determinable), the results of the comparison, and any additional procedures performed on significant unexpected differences with their results. Recording the acceptable threshold is strong practice because it shows how differences were evaluated:

+-------------------------------------------------------------------------------------------------------+
|                           AU-C 520 DOCUMENTATION (3 REQUIRED ITEMS + THRESHOLD)                               |
|                                                                                                       |
|  1. THE EXPECTATION:                                                                                  |
|     The auditor must document the expectation developed and the specific factors considered in its   |
|     development (including data sources, assumptions, and formulas).                                  |
|                                                                                                       |
|  2. THE COMPARISON RESULTS:                                                                           |
|     The recorded amounts, the expected values, and the quantitative comparison between the two.       |
|                                                                                                       |
|  3. THE ACCEPTABLE THRESHOLD (good practice):                                                                         |
|     The maximum difference acceptable without further investigation (tolerable difference).           |
|                                                                                                       |
|  4. INVESTIGATION OF UNEXPECTED DIFFERENCES:                                                          |
|     Any additional auditing procedures performed when differences exceeded the threshold, management's|
|     explanations, and the specific CORROBORATING AUDIT EVIDENCE obtained.                            |
+-------------------------------------------------------------------------------------------------------+

---

5. Concrete Substantive Analytical Models

Model A: Apartment Complex Rental Revenue Reasonableness Test

• Client Facts: Horizon Properties owns a 200-unit residential apartment building. Rent is fixed at $1,800/month per unit. The certified property manager's logs show an average occupancy rate of 95% for the 12-month fiscal year. The client recorded rental revenue of $4,350,000.
• Auditor's Independent Expectation: $\text{Expected Revenue} = 200\text{ units} \times 95\%\text{ occupancy} \times \$1,800/\text{month} \times 12\text{ months} = \$4,104,000$
• Auditor's Tolerable Difference Threshold: $50,000.
• Variance Analysis: $\text{Difference} = \$4,350,000\text{ (Recorded)} - \$4,104,000\text{ (Expected)} = \$246,000$
• Audit Action: The variance of $246,000 far exceeds the tolerable threshold of $50,000. The auditor cannot accept the recorded amount. The auditor must inquire of management and independently corroborate any explanation (e.g., verifying whether unearned upfront lease prepayments were improperly recognized in revenue, which would violate cutoff and occurrence).

Model B: Long-Term Debt Interest Expense Reasonableness Test

• Client Facts: Apex Corporation has three fixed-rate term bonds outstanding throughout the entire year:
  1. Tranche A: $10,000,000 at 5.0% coupon ($500,000)
  2. Tranche B: $15,000,000 at 6.0% coupon ($900,000)
  3. Tranche C: $5,000,000 at 4.0% issued on July 1 (half-year = $100,000)
• Auditor's Independent Expectation: $1,500,000.
• Client Recorded Expense: $1,502,500.
• Variance: $2,500, which is well within the tolerable difference of $25,000. Substantive conclusion: Interest expense is materially accurate at the assertion level.