7032 Suitability of analytical procedures

Jul-2017

OAG Guidance

In determining the suitability of substantive analytical procedures given the assertions, consider:

a) The assessment of the risk of material misstatement

The following are important considerations in determining the risk of material misstatement:

• understanding of the entity and its system of internal control,
• magnitude and likelihood of misstatement of the items involved, and
• nature of the assertionassertion.

For example, if we have no or limited evidence from controls over sales order processing and we will have to derive more assurance from tests of details rather than substantive analytical procedures for assertions related to receivables, substantive analytical procedures would provide low, if any, assurance. As another example, when inventory is a significant FSLI, ordinarily we would not rely only on substantive analytical procedures when performing audit procedures on the valuation assertion which tends to have a higher likelihood of misstatement.

For further guidance on developing our testing strategy, see OAG Audit 7010.

b) Any tests of details directed toward the same assertion

Substantive analytical procedures may be considered appropriate when tests of details are performed on the same assertion. Inherent in these considerations are factors that affect the assurance and precision of the substantive analytical procedure as discussed in OAG Audit 7033.1 (disaggregation, reliability, predictability, type of analytical procedure, etc.). For example, if source data for a substantive analytical procedure is not reliable or if an account balance is highly volatile, then assurance that can be derived from substantive analytical procedures may be significantly reduced or even not achievable.

Types of analytical procedures

Choose among the following five types of analytical procedures may be appropriate based on our objectives for the procedure (i.e., purpose of the test, desired level of assurance):

1. Trend analysis—the analysis of changes in an account over time.

2. Ratio analysis—the comparison, across time or to a benchmark, of relationships between financial statement accounts and between an account and non-financial data.

3. Reasonableness testing—the analysis of accounts, or changes in accounts between accounting periods, that involves the development of a model to form an expectation based on financial data, non-financial data, or both.

4. Regression analysis—the use of statistical models to quantify our expectation, with measurable risk and precision levels.

5. Scanning analytics—the identification of anomalous individual items within account balances or other entity data through the scanning or analysis of entries in transaction listings, subsidiary ledgers, general ledger control accounts, adjusting entries, suspense accounts, reconciliations, and other detailed reports. Scanning analytics always has to be combined with appropriate tests of details following up on the anomalous items.

Each of the five types uses a different method to form an expectation. The five types are defined below, with the first four ranked from lowest to highest in order of their inherent precision. Scanning analytics are different from the other types of analytical procedures in that scanning analytics search within accounts or other entity data to identify anomalous individual items, while the other types use aggregated financial information.

Following this is guidance on factors that affect these types of analytical procedures that are considered during the design and execution of the procedures.

Benchmarking can also be considered for use in analytical procedures as it can improve the rigor of the procedure. See OAG Audit 7035 for further information.

The following basic financial statement analyses using trend and ratio analysis may be useful for risk assessment, overall conclusion and substantive analytical procedures:

1. Comparative financial statement analysis—Side-by-side comparisons of two or more periods of a financial statement to analyze trends.  Trends can be analyzed across time and to benchmark information.

2. Common size financial statement analysis—Converting financial statement amounts into percentages of a related total, such as sales for income statement items or either sales or total assets for balance sheet items, and performing comparisons across time or to benchmark information to analyze the internal structure of the financial statements. For more information on benchmarking, please see OAG Audit 7035.

OAG Guidance

Trend analysis

This is the analysis of changes in an account over time. Simple trend or fluctuation analyses typically compare a prior period’s account balance with the current balance. Trend analysis can also encompass multiple time periods and includes comparing recorded trends with budget amounts and benchmarking against competitor and industry information.

Trend analysis is most appropriate when the account or relationship is fairly predictable (e.g., sales or equipment in a stable environment). It is less effective when the entity has experienced significant changes. The number of time periods used in trend analysis is a function of the stability of operations. The more stable the operations over time, the more predictable the relationship and the more appropriate the use of multiple periods. Using more periods can increase the precision of the expectation.

The use of only the prior period balance as the expectation in trend analysis reduces the effectiveness of the analytical procedure. Accordingly, where operations are stable, more than two periods are normally included in the analysis. When we incorporate more than two periods in our trend analyses, it means we examine the past trends to develop an expectation for the current period. For example, if sales have steadily increased between four and six percent in each of the last five years, our expectation for sales in the current year when performing a trend analysis might be five percent greater than the prior year.

Trend analysis at an aggregated level (e.g., on a consolidated basis) is relatively imprecise because a significant difference will often be small relative to the natural variation (i.e. regular fluctuation of the balance) in an aggregated account balance. This suggests the need to perform trend analysis on a disaggregated level (e.g., by segment, product, location, or quarterly or monthly) where greater precision is needed, such as when performing substantive analytical procedures.

Trend analysis can be performed using comparative financial statement analysis, where financial statement information is compared side-by-side, and/or common size financial statement analysis, where all financial statement components are converted to a percentage of one component, e.g., a percentage of sales or total assets. For a further description of comparative and common size financial statement analysis, see OAG Audit 7035.

Because trend analysis relies on a single predictor (e.g., prior periods’ data for an account balance), it does not incorporate the use of other potentially relevant financial and operating data, as the other types of analytical procedures can do. Trend analysis performed on lead schedules prepared for FSLIs may not typically be designed with the appropriate rigor and precision to provide substantive audit evidence akin to substantive analytical procedures, however we may prepare lead schedules following the four-step process if we seek audit evidence from such procedures.

Benchmarking perspective:

All financial benchmarks could be relevant to Trend Analysis. Select ranges of key benchmarks and create a trend analysis, e.g., benchmark revenues and gross margins against the industry over a period of time (monthly, quarterly or yearly) to detect unusual trends relative to the industry. See OAG Audit 7035 for further guidance.

Ratio analysis

This is the comparison, across time or to an external benchmark, of relationships between financial statement accounts (e.g., return on equity) or between an account and non-financial data (e.g., cost per order or sales per square foot). Benchmarks for ratio analysis typically involve competitor or industry data (e.g., comparing the entity’s gross profit margin with its competitors or industry aggregated data). When comparing company ratios across time, an entity’s prior period performance is used as a benchmark.

Ratio analysis also includes common size analysis, where all financial statement components are converted to a percentage of one component, e.g., a percentage of sales or total assets. For a further description of common size analysis, see OAG Audit 7035.

Like trend analysis, ratio analysis is more appropriate when the relationship between accounts is fairly predictable and stable (e.g., the relationship between sales and accounts receivable). Ratio analysis can be more effective than trend analysis because comparisons between accounts can often reveal unusual fluctuations that analysis of the individual accounts would not.

Benchmarking ratios with competitors or aggregated industry data are most useful when operating factors are comparable. Also like trend analysis, ratio analysis at an aggregated level (i.e., consolidated operating units or across product lines) is relatively imprecise because a significant difference is often small relative to the natural variation in the ratios. Thus, when increased effectiveness is desired, perform ratio analysis on a disaggregated level (e.g., by segment, product, or location).

See OAG Audit 7035 for a listing of common ratios used in ratio analysis. Also use ratios that are relevant to a specific entity and industry (e.g., revenue per click-through, revenue per barrel, cost per student).

Benchmarking perspective:

Ratios, such as liquidity, activity, profitability could be benchmarked against competitors or industry aggregated data to help us detect risk. For example, comparison of sales and accounts receivable (e.g., Days Sales Outstanding or accounts receivable as a percentage of sales) over time, compared to competitors or industry data, could indicate problems with revenue recognition if sales are growing faster than accounts receivable or, vice versa, valuation of accounts receivable if accounts receivable is growing faster than sales.

Reasonableness testing

This is the analysis of account balances, or changes in account balances between accounting periods, that involves the development of a model to form an expectation based on financial data, non-financial data, or both. In many cases, a simple model may be sufficient. For example, an expectation for hotel revenue may be developed using a model that includes the average occupancy rate and the average room rate by category or class of room. Similarly, to develop an expectation for payroll expense, we may use a model including number of employees, pay rates, hire and termination dates, and overtime.

Reasonableness testing relies on our knowledge of the relationships and factors that affect the account balances. Use that knowledge to develop assumptions for each of the key factors (e.g., industry and economic factors) to estimate the account balance. A reasonableness test for sales could be formed by considering the number of units sold, the unit price by product line, different pricing structures, and an understanding of industry trends during the period.

Regression analysis

This is the use of statistical models to quantify our expectation, with measurable risk and precision levels. For example, we might form an expectation for sales by performing a regression analysis using model inputs such as management’s sales forecast, commission expense (if it is not a fixed percentage of sales), and advertising expenditures.

Standard software packages, including Microsoft Excel, can be used to perform regression analysis.

Regression analysis is similar to reasonableness testing in that the model yields a specific explicit prediction using inputs based on our knowledge of the factors that affect the account balances. Regression analysis is most effective when data is disaggregated, from a system with good internal control, and has a strong predictable relationship between at least two data elements (e.g., retail sales for a clothing store and the square meters of the store).

Scanning analytics

This is the identification of anomalous individual items within account balances or other entity data through the scanning or analysis of entries in transaction listings, subsidiary ledgers, general ledger accounts, adjusting entries, suspense accounts, reconciliations, and other detailed reports.

Scanning analytics are different from other types of analytical procedures, which are used at a more macro level with a top-down focus. Scanning analytics include searching for large or unusual or unexpected items in the accounting records (e.g., non-standard journal entries), as well as reviewing transaction data (e.g., suspense accounts, adjusting journal entries) for indications of errors that have occurred or to identify items that may be more prone to error.

Scanning analytics may include the following types of tests (listing is not all-inclusive):

• Unusual items—scanning transaction listings or closing/adjusting entries or large and unusual items.

• Gaps—identifying missing items by reference to sequential numbering or other consistent patterns as in reviews of check registers or sales ledgers for missing items in the sequence.

• Duplicates— identifying duplicate invoice numbers, payments, or payroll transactions to the same payee.

• Filters—applying or comparing patterns to identify only items of interest as in identifying all new suppliers, non-standard journal entries, accounts under dispute, transactions with related parties, inactive inventory, raw materials related to obsolete inventory, or collections after a period-end for specific accounts.

• Sorts—categorizing data to identify specific items by e.g., sorting suspense account items in reverse date order or identifying all payments to a specific vendor.

• Statistics—identifying items by specifying statistical differences such as identifying all vendor payments or unit prices more than two standard deviations from the mean.

• Aging—ordering items by reference to date such as identifying all invoices over 90 days past due or identifying old inventory.

• Classifying—identifying items by reference to a characteristic in such as identifying credit balances in accounts receivable.

• Stratifying—categorizing data by reference to a characteristic in the population, e.g., grouping customer accounts by balance size or employee by overtime pay.

• Comparison—comparing data by reference to a second source as in comparing customer sales contracts per the sales report to data in a sales tracking system for identification of duplicate, missing or different items.

The expectation formed for scanning analytics depends on the purpose of the procedure. The expectation in scanning for large or unusual or unexpected items is based on our assessment of what constitutes “normal” or “expected.”

While some scanning analytics may be performed manually (e.g., scanning closing or adjusting entries), others (e.g., filters, duplicates, gaps, sorts) may require computer-assisted audit techniques using software packages like IDEA. Specialists can assist in the development and execution of scanning analytics using more complex computer-assisted audit techniques.

Suitable tests of details need to be designed to follow up on the anomalous items identified, in order to obtain sufficient appropriate audit evidence as to whether there is a misstatement.

Summary

All five types of analytical procedures can be used at any time in the audit; and the five types are not mutually exclusive, meaning we can do a trend analysis of ratios; however, certain types are more commonly used for one purpose than for the others (i.e., risk assessment, substantive, or overall conclusion). When to use the various types of analytical procedures is discussed in more detail as part of the discussion of each purpose.

Trend and ratio analyses are less precise than reasonableness testing or regression analysis. To improve the precision of a trend or ratio analysis, use highly reliable data (internal entity data that has been subjected to control testing or external data such as peer data that has been audited or industry data from trade associations).

Reasonableness tests or regression analyses are more precise, but take care to verify the nature of the relationship being tested is equally precise (e.g., expected annual interest expense computed for aggregate debt where the debt portfolio contains different floating-rate instruments in a time of changing interest rates is not as precise as using the individual debt instruments).

The appropriate level of precision needed for a particular analytical procedure is a matter of professional judgment. Expectations that are insufficiently precise for their intended purpose can reduce both effectiveness and efficiency. For example, the effectiveness of substantive analytical procedures is reduced when the expectation is less able to highlight potential material misstatements (perhaps due to offsetting material misstatements in high-level account balances that mask true differences) or provide the necessary level of assurance. Imprecise and uninformed expectations can also reduce efficiency by yielding “false” differences requiring follow-up and additional documentation.

In summary, there is a direct correlation between the type of analytical procedure selected and the precision it can provide. Generally, the more precision inherent in an analytical procedure, the greater is the potential reliability of that procedure.

OAG Guidance

Two types of basic financial statement analyses that may be useful as a stand-alone analytical procedure or in conjunction with other analytical procedures include:

• comparative financial statement analysis, and
• common size financial statement analysis.

Comparative financial statement analysis

Comparative financial statement analysis, or horizontal analysis, results from side-by-side comparisons of two or more periods of a financial statement followed by analysis of the changes (flux analysis). This analysis is performed to identify trends in the financial statements. The benefit of this analysis is the additional information provided by understanding the direction, rate, and volatility of a trend and its relationship to correlating trends. The more periods included in the trend, the more beneficial the analysis would generally be (e.g., a five-year trend or an eight-quarter trend).

Benchmarking and comparative financial statements

Comparative financial statements are useful for benchmarking a company to its competitors or industry statistics to analyze differences in trends. By comparing our knowledge of the business and trends with industry and competitor information a greatly enhanced analytical procedure for assessing risk and improving audit evidence will result. Benchmarking can be considered as an addition to analytical procedures performed on financial information. Please see OAG Audit 7035 for more information on using benchmarking.

Common-size financial statement analysis

Common-size financial statement analysis, or vertical analysis, results from converting financial statement amounts into percentages of a related total. This analysis is performed to enhance the analysis of the internal structure of the financial statements. The benefit of this analysis is the additional information provided by understanding the proportion of lines items of a related total.

Examples:

• The balance sheet is generally common-sized by expressing individual line items on the balance sheet as a percentage of either total assets or sales.

• The profit and loss statement is generally common-sized by expressing all line items below the sales line item as a percentage of sales.

OAG Guidance

Typically, test of details are often the most appropriate way to test compliance with authorities.

Although many types of analytical procedures can be used in a particular audit, some are more suitable for public sector auditing. For instance, trend analysis, reasonableness testing, and scanning analytics are more often used, while ratio analysis and regression analysis are rarely used in public sector auditing.

Considering the specificity of each governmental entity and of its activities, benchmarking with other entities within the government is rarely used in the context of an annual audit.