Flux and Variance Analysis: A Practical Guide for Close and Audit

Flux analysis (also called variance analysis) compares account balances or activity between periods, or against budget, and investigates the changes that matter. Controllers use it as a close control to catch posting errors before statements go out. Auditors use the same comparison as a risk assessment tool at planning, as a substantive analytical procedure when an expectation is built properly, and as an overall review near the end of the file.

The comparison itself is the easy part. The value comes from the threshold, the expectation and the quality of the explanation. This guide covers all three, with a worked example and the mistakes that make a flux schedule useless.

Two uses, one schedule

An auditor can read the client's flux file to understand the business, but the client's explanations are not audit evidence until the auditor corroborates them. See analytical procedures under AU-C 520.

Setting thresholds

A flux threshold usually has a dollar test and a percentage test. Which one controls is a policy decision, and a common rule is to investigate when both are exceeded, because dollar alone flags big accounts with small changes and percent alone flags small accounts with large changes.

• Close threshold example: investigate where the change is at least $25,000 and at least 10 percent.
• Audit threshold: tie it to performance materiality, so a difference that could hide a material misstatement is investigated. Use the materiality calculator for the base figures. AU-C 520 requires the auditor to determine the difference from expectation that can be accepted without further investigation.

Review the threshold annually and when the entity's size changes. A threshold set for a $2 million entity does not work after the entity reaches $20 million.

Building the expectation

For close, the prior period is the usual baseline. For audit evidence, build a better expectation from independent data. Examples:

• Payroll: headcount by month times average rate, plus employer taxes.
• Rent: lease schedule.
• Interest expense: average debt balance times the contractual rate.
• Revenue: units shipped times price, or prior revenue adjusted for known price and volume changes.
• Cost of goods sold: expected from prior-year margin applied to current revenue.

An expectation is only as good as its data. Use data the client did not generate from the account being tested, or test its reliability.

Worked example

All numbers are illustrative. Thresholds are at least $25,000 and at least 10 percent, both required. Prior-year revenue was $2,400,000 and current-year revenue is $2,640,000.

The percentages: 240,000 / 2,400,000 = 10.0%; 276,000 / 1,440,000 = 19.2%; 38,000 / 40,000 = 95.0%.

Gross margin fell from (2,400,000 - 1,440,000) / 2,400,000 = 40.0% to (2,640,000 - 1,716,000) / 2,640,000 = 35.0%.

Expectation for cost of goods sold

If revenue grew 10 percent and margin were stable, cost of goods sold would be expected at 1,440,000 x 1.10 = 1,584,000. The recorded amount is 1,716,000, a difference of 132,000 (8.3 percent of the expectation). That is the figure to run down, not the 276,000 year-over-year change. Candidate explanations to test: supplier price increases (check vendor invoices and price lists), a mix shift to lower-margin products (check sales by product), inventory costing errors or a cutoff error (check receiving and shipping records), and shrinkage or write-downs (check the count and reserve entries).

Two items below the line

Bad debt expense is up 158.3 percent, but the dollar change is below the threshold. A flux policy with a dollar-and-percent rule would not flag it, even though it signals collection trouble. A reviewer should read the lines below the threshold for direction and compare them to related accounts, here receivables aging. See the receivables aging and allowance guide. Similarly, a 66.7 percent increase in interest expense deserves a quick reasonableness check against the debt schedule.

Writing explanations that survive review

A weak explanation restates the number or blames timing. A strong one names the driver, quantifies it and points to evidence.

Each explanation should answer: what changed, by how much, why, and how we know. If explanations do not add up to the variance, the unexplained remainder is the finding.

Process for the close

1. Run the comparison from the trial balance after adjusting entries.
2. Apply the threshold and flag items.
3. Assign each flagged item to the account owner for a written explanation.
4. Reviewer challenges explanations and requests support for those that matter.
5. Post corrections, then rerun.
6. File the schedule and sign-offs with the close binder. See the month-end close checklist.

The flux analysis calculator computes the dollar and percent changes and flags items against thresholds.

Common mistakes

• Using a single threshold for every account, including accounts that should not move.
• Leaving explanations to the person who made the entry.
• Accepting "timing" without naming the entry.
• Comparing to prior year when the business changed: acquisitions, new products, lost customers.
• Treating the client's flux explanations as audit evidence without corroboration.
• Not flagging items that should have changed but did not, such as depreciation after an asset purchase.

FAQ

What is flux analysis?

A comparison of financial statement lines between periods or against budget, followed by investigation of the significant changes.

What threshold should I use?

Set a dollar and a percentage threshold tied to materiality and to the entity's size. For audit use, base it on performance materiality and the precision of the expectation.

Is flux analysis audit evidence?

Only when built as a substantive analytical procedure under AU-C 520 with a sufficiently precise expectation, reliable data and corroborated explanations. A year-over-year comparison alone is a risk assessment tool.

How is flux analysis different from variance analysis?

The terms are often used interchangeably. Flux commonly refers to period-over-period changes in actuals, and variance to differences from budget or standard.

How often should flux analysis be performed?

Monthly at close for key accounts, with a fuller review at year end.

Sources

• AICPA, AU-C section 520, Analytical Procedures (requirements for substantive analytical procedures: expectation, data reliability and investigation threshold) and AU-C section 315
• AICPA, AU-C section 500 and AU-C section 330 on evidence and further audit procedures

AU-C references are to AICPA Professional Standards, paraphrased here; read the standard itself before relying on a requirement. This page is guidance, not professional advice. Report an error