Analytical Procedures
Analytical procedures involve evaluating financial information through the study of plausible relationships among both financial and nonfinancial data. They are one of the most versatile tools available to auditors—used in planning, substantive testing, and the final review of the financial statements.
Definition
Analytical procedures consist of evaluations of financial information made by studying plausible relationships among financial and nonfinancial data. They involve comparisons of recorded amounts (or ratios derived from them) to expectations developed by the auditor. When recorded amounts deviate significantly from expectations, the auditor investigates the cause.
Common forms of analytical procedures include:
- Trend analysis — Comparing current-year balances to prior periods
- Ratio analysis — Calculating financial ratios and comparing them to expectations
- Reasonableness testing — Developing an independent estimate and comparing it to the recorded balance
- Regression analysis — Using statistical models to predict expected values
When Are Analytical Procedures Required vs. Optional?
Understanding when analytical procedures are mandatory versus discretionary is critical for the CPA exam.
| Audit Phase | Required? | Purpose |
|---|---|---|
| Planning | ✅ Required | Identify areas of potential misstatement; enhance the auditor's understanding of the entity and its environment |
| Substantive Testing | ❌ Optional | May be used as a substantive procedure to detect material misstatements at the assertion level |
| Overall Review (Near Completion) | ✅ Required | Assess whether the financial statements as a whole are consistent with the auditor's understanding of the entity |
Analytical procedures are required during planning and during the overall review near the end of the audit. They are optional during substantive testing. The exam frequently tests whether candidates know this distinction.
Planning Phase
During planning, analytical procedures help the auditor identify unusual transactions, unexpected balances, and areas where the risk of material misstatement may be elevated.
While planning the audit of Gies Co., the auditor compares the current year's gross margin to the prior three years. A significant decline triggers further investigation and may lead the auditor to classify revenue recognition as a higher-risk area.
Overall Review Stage
At the end of the audit, analytical procedures serve as a "sanity check." The auditor steps back and considers whether the financial statements, taken as a whole, are consistent with the auditor's understanding of the entity developed throughout the engagement.
- Do the financial statements make sense given the auditor's knowledge of the industry, the economy, and the client's operations?
- Are there any previously unrecognized risks of material misstatement?
If the overall review reveals unexpected relationships, the auditor may need to perform additional procedures before issuing the opinion.
Using Analytical Procedures as Substantive Tests
When used as substantive procedures, analytical procedures must meet a higher standard of precision and rigor. The auditor follows a structured process:
Step 1: Determine Suitability
Evaluate whether an analytical procedure is appropriate for the assertion being tested. Analytical procedures work best when:
- The data relationships are predictable and stable
- The account balance can be reasonably estimated from available data
Analytical procedures are generally more suitable for income statement accounts (which accumulate over time and are more predictable) than for balance sheet accounts (which represent a point-in-time snapshot and may be more volatile).
Step 2: Evaluate Reliability of Data
The auditor must assess whether the data used to develop the expectation is reliable. Consider:
- The source of the data (internal vs. external)
- Whether the data was subjected to audit testing
- The conditions under which the data was gathered
Step 3: Develop an Expectation
The auditor forms an independent expectation of the recorded amount. The more precise the expectation, the more effective the procedure.
To test Illini Security's total revenue, the auditor gathers data on the number of active client contracts, average monthly contract value, and the number of months in the period. Multiplying these factors produces an independent estimate of expected revenue.
Step 4: Compare to Recorded Amount
The auditor compares the expected amount to the client's recorded balance and determines whether the difference exceeds a pre-established threshold of acceptable variation.
Step 5: Investigate Significant Differences
If the recorded amount differs materially from the expectation, the auditor:
- Inquires of management to obtain explanations
- Corroborates management's explanations with additional evidence
- Performs alternative procedures if the explanation is insufficient
Audit Data Analytics (ADAs)
Audit data analytics represent the application of data analysis techniques—often technology-driven—to audit procedures. ADAs enhance the auditor's ability to identify patterns, anomalies, and risks in large datasets.
Definition
Audit data analytics (ADAs) are techniques that analyze complete populations or large datasets of financial and nonfinancial information to identify patterns, correlations, and anomalies relevant to the audit.
Benefits of ADAs
- Analyze entire populations rather than samples
- Identify unusual transactions that might escape traditional sampling
- Improve audit quality through deeper understanding of the data
- Enhance efficiency by focusing attention on high-risk items
Steps for Implementing ADAs
- Define the objective — Determine what the auditor is trying to accomplish
- Obtain and prepare data — Extract data from the client's systems and prepare it for analysis
- Perform the analysis — Apply the chosen analytical technique
- Interpret results — Evaluate findings in the context of the audit
- Document conclusions — Record the procedures performed and conclusions reached
Four Categories of Audit Data Analytics
| Category | Question It Answers | Example |
|---|---|---|
| Descriptive | What happened? | Summarize all journal entries posted by MAS Inc.'s accounting staff during the year, showing volume and dollar amounts by user |
| Diagnostic | Why did it happen? | Investigate why Kingfisher Industries' warranty expense spiked in Q3 by correlating with product return data |
| Predictive | What is likely to happen? | Forecast expected bad debt expense for BIF Partners based on historical collection patterns and current aging |
| Prescriptive | What should we do about it? | Recommend which inventory locations at MSA Records should receive additional testing based on historical discrepancy rates |
Descriptive and diagnostic analytics are most commonly used in current audit practice. Predictive and prescriptive analytics represent the frontier of audit innovation and are increasingly tested on the CPA exam.
ADA Applications
Applied to Tests of Details
ADAs can enhance tests of details by enabling the auditor to:
- Analyze all transactions in a population rather than relying on a sample
- Flag transactions with unusual characteristics (e.g., entries posted outside business hours, round-dollar amounts, entries just below authorization thresholds)
- Match datasets to identify exceptions (e.g., comparing vendor addresses to employee addresses to detect fictitious vendors)
The auditor uses data analytics to compare Illini Entertainment's complete vendor master file against the employee address database. Three matches are found, prompting further investigation into potential fictitious vendor fraud.
Applied to Analytical Procedures
ADAs can strengthen analytical procedures by:
- Incorporating more data points and variables into the analysis
- Using visualization tools to identify trends and outliers
- Building statistical models that produce more precise expectations
Methods to Obtain ADA Data
Auditors can obtain data for analytics through several methods:
- Direct system access — Connecting to the client's ERP or accounting system
- Data exports — Requesting structured data files (CSV, Excel, XML) from the client
- API connections — Pulling data programmatically from the client's systems
- Third-party data — Obtaining external benchmarking or industry data
- Screen scraping / OCR — Extracting data from documents or legacy systems
The auditor must evaluate the completeness and accuracy of data obtained for analytics. If the underlying data is unreliable, the ADA results will be misleading. This is analogous to evaluating the reliability of data used in any analytical procedure.
Techniques to Interpret ADA Results
Interpreting the output of audit data analytics requires both technical and professional skills:
- Filtering and sorting — Isolating items that meet specific criteria (e.g., all transactions over a threshold)
- Statistical analysis — Applying measures of central tendency, dispersion, and correlation
- Pattern recognition — Identifying recurring anomalies or clusters of unusual activity
- Benchmarking — Comparing results to industry data, prior periods, or budgets
- Exception testing — Focusing on items that fall outside expected parameters
Data Visualizations
Visual representations of data can make patterns and anomalies easier to identify. Common visualizations used in audit analytics include:
| Visualization | Use Case |
|---|---|
| Bar / column charts | Comparing balances across periods or categories |
| Line charts | Identifying trends over time (e.g., monthly revenue at Gies Co.) |
| Scatter plots | Detecting outliers in large datasets |
| Heat maps | Highlighting areas of concentration or risk |
| Histograms | Showing the distribution of transaction amounts |
| Pie charts | Illustrating proportional composition (e.g., revenue by segment at Illini Entertainment) |
On the CPA exam, you may be asked to identify which type of visualization is most appropriate for a given audit objective. Scatter plots are ideal for outlier detection, while trend lines are best for time-series comparisons.
Relational Databases
Much of the data used in audit analytics resides in relational databases—structured systems that organize data into related tables.
Key concepts:
- Tables — Store data in rows (records) and columns (fields)
- Primary key — A unique identifier for each record in a table
- Foreign key — A field in one table that references the primary key of another table, establishing a relationship
- Queries — Requests to retrieve, filter, or combine data from one or more tables (commonly written in SQL)
Kingfisher Industries stores its accounting data in a relational database. The auditor writes a query joining the sales order table, shipping table, and invoice table to identify shipments that were never invoiced—a test of the completeness assertion for revenue.
The ETL Process (Extract, Transform, Load)
When performing audit data analytics, auditors frequently work with data from multiple sources. The ETL process standardizes this workflow:
Extract
Pull raw data from the client's systems, databases, or external sources. This may involve:
- Downloading data files
- Running database queries
- Accessing data through APIs
Transform
Clean and standardize the extracted data to make it suitable for analysis:
- Removing duplicates
- Standardizing date formats, currency codes, and naming conventions
- Validating data integrity (e.g., checking for missing values or impossible entries)
- Converting data types as needed
Load
Import the transformed data into the auditor's analytical tool or platform for analysis.
The ETL process is essential for ensuring that the data used in analytics is complete, accurate, and in a usable format. Errors in any stage—extraction, transformation, or loading—can compromise the reliability of the entire analysis.
| Stage | Key Activity | Risk If Done Poorly |
|---|---|---|
| Extract | Pull data from source systems | Incomplete data; missing records |
| Transform | Clean, standardize, and validate data | Inaccurate analysis due to data quality issues |
| Load | Import into analysis tool | Corruption during import; format incompatibility |