How to Audit Accounts Receivable: Confirmations, Cutoff, and CECL

To audit accounts receivable, you answer four questions with evidence: do the receivables exist, will the company collect them, are any missing from the books, and are they recorded in the right period. Everything else in the engagement, from confirmations to cutoff testing to the allowance review, exists to answer one of those four. Because receivables often make up a large share of current assets, even a modest misstatement can distort revenue, overstate liquidity, and mislead lenders and investors, so the work has to be built carefully from the ground up.

What follows walks through the procedures in the order you’d actually perform them, from pulling documents to writing up findings.

Start With the Documents You Need

Ask the client for the accounts receivable trial balance first. This is a customer-by-customer listing of every outstanding balance as of the audit date, and it must tie exactly to the single control account balance in the general ledger. If the subsidiary detail and the control account don’t match, stop and resolve the discrepancy before doing anything else. Substantive testing built on a ledger that doesn’t reconcile to its own summary is wasted effort.

Next, get the aging report. It groups each customer’s balance by how long it has been outstanding, typically in current, 31–60 days, 61–90 days, and over 90 days past due. The aging drives two later steps: it guides sample selection for confirmations, and it’s the primary tool for evaluating whether management’s estimate of uncollectible amounts is reasonable. Confirm the aging is calculated from the original invoice date rather than the statement date; the wrong starting point skews every bucket and hides how stale balances really are.

Round out the request list with credit policies and approval documentation, a schedule of write-offs during the year, the prior-year workpapers for comparison, and bank statements covering the period just after year-end for cutoff testing.

Test the Controls Over the Credit-to-Collection Cycle

Before touching individual balances, evaluate whether the company’s controls over receivables actually work. Weak controls increase the risk that errors or fraud have gone undetected, and that pushes you toward larger samples and heavier substantive procedures.

The controls that matter most:

Credit approval by someone independent of sales. If salespeople can extend credit on their own, the company is inviting bad receivables.
Segregation of duties between cash application, credit memo issuance, and write-offs. Combining these functions creates an easy path for fraud, especially lapping.
Invoice-to-shipment matching. Every invoice should trace to a shipping document; investigate anything that doesn’t.
Cash application review. Trace a sample of incoming payments from bank deposit to customer account to confirm cash is applied to the correct invoices.
Monthly reconciliation of the subsidiary ledger to the control account, with follow-up on discrepancies.
Review of any manual journal entries hitting A/R. These bypass normal billing and collection, so each one deserves scrutiny.

Walk through each control, pick a sample of transactions from the period, and test whether the control actually operated as designed. If controls aren’t working, adjust the audit plan to compensate.

Confirm Balances Directly With Customers

External confirmation is the strongest evidence available for whether a receivable exists and belongs to the company. You send the request directly to the customer, and the reply comes straight back to you without touching management. That independence is what makes confirmations persuasive.

Use stratified sampling: select all large-dollar accounts and a random sample of smaller ones. Two formats are available. A positive confirmation asks the customer to respond either way, either agreeing with the balance or explaining the disagreement. Some auditors use a blank version that omits the dollar amount and asks the customer to fill it in, which yields more reliable evidence but usually a lower response rate. Positive confirmations are the default for large balances and for any situation with identified control weaknesses.¹

A negative confirmation asks the customer to respond only if they disagree, treating silence as agreement. This format is only appropriate when three conditions are met: the risk of material misstatement is low, controls have been tested and found effective, and the population consists of many small, homogeneous balances with a low expected exception rate. Negative confirmations alone never provide sufficient evidence on their own, so use them alongside other substantive procedures.¹

When Customers Don’t Respond

Non-responses are routine. When a positive confirmation comes back blank, follow up with the customer first. If that produces nothing, perform alternative procedures.

The most effective alternative is examining subsequent cash receipts. If the customer paid the balance shortly after year-end, that payment is strong evidence the debt was real and collectible. Match the payment to the specific invoices being paid rather than to the total, since a payment on new invoices doesn’t prove the old balance existed.¹

When no subsequent payment has come in, work through the supporting documents: the original sales invoice, the customer’s purchase order, the signed contract, and shipping records such as a bill of lading. Together they show the transaction happened, goods left the warehouse, and the company has a right to payment. If a sampled item still can’t be verified, expand the sample. A pattern of unverifiable balances is a red flag that calls for broader testing.

Evaluate the Allowance for Credit Losses

Confirming that receivables exist is only half of valuation. The other half is whether the company will actually collect them. Every company holding receivables must estimate the portion it expects to lose and record that estimate as an allowance for credit losses, which brings the receivable down to its net realizable value on the balance sheet.

What CECL Requires

Under the current expected credit losses model, companies must estimate lifetime expected losses from the moment a receivable is recorded, not when a loss becomes probable. Historical loss data still matters but is no longer sufficient on its own. The estimate must also reflect current conditions and reasonable, supportable forecasts of future economic conditions.²

Many companies still use aging-based provision matrices, applying loss percentages to each age bucket. That approach is acceptable under CECL, but the loss rates have to be adjusted for forward-looking information. If the economy is deteriorating or a major customer’s industry is contracting, historical rates alone will understate the allowance. When you review the estimate, check whether management has segmented receivables into pools with similar risk characteristics (by product line, industry, or geography, for example) and whether the rates for each pool reflect more than backward-looking data.

A 2025 update introduced a practical expedient for current trade receivables. Companies may assume that current conditions as of the balance sheet date remain unchanged for the remaining life of the receivable.²

Testing the Estimate

Start by analyzing three to five years of write-off history. Calculate the percentage of credit sales that proved uncollectible in each year and compare that trend to the percentage management is applying now. A sudden drop in the estimated rate without a matching improvement in collections suggests the allowance is understated.

Scrutinize old balances on the aging. Receivables past due more than 90 or 120 days are far less likely to be collected, and the loss percentages applied to them should be significantly higher than those for current balances. Look for specific accounts that are clearly impaired: customers in bankruptcy, customers who have gone silent despite repeated collection efforts, or customers in active disputes. These should be assessed individually rather than lumped into a general percentage.

If the allowance is too low, propose an adjustment that increases bad debt expense and reduces the net receivable. An allowance that’s too high also needs correcting; overstating the reserve depresses current earnings and gives management a cushion to release in a later period to smooth results.

Test Cutoff and Completeness

Cutoff errors are one of the easiest ways to manipulate revenue and often the hardest to spot without targeted testing. The goal is to verify every transaction landed in the correct period.

Sales Cutoff

Pull the last several sales invoices recorded before year-end and the first several recorded in the new year. For each invoice, trace the date to the corresponding shipping document. Under current revenue recognition standards, a sale is recorded when the company satisfies its performance obligation, which for a standard product shipment usually means when control of the goods transfers to the customer.³

If a December 31 invoice is supported by a January 2 shipping document, the sale was booked too early. That error overstates both revenue and receivables and requires a correcting entry. “Upon shipment” is not always the trigger. Under ASC 606, recognition depends on the contract terms; some contracts transfer control on delivery rather than shipment, and some obligations are satisfied over time. When you hit a complex arrangement, read the contract and evaluate when control actually transfers.

Cash Receipts Cutoff

The same logic runs in reverse for cash coming in. Payments recorded on December 31 have to be verified against bank deposit records to confirm the funds were actually received by year-end. A payment that arrived on January 2 but was backdated to December 31 understates the year-end receivable and overstates cash.

Completeness

Completeness runs opposite to existence. Instead of starting from the books and searching for proof, start from the source documents and confirm they made it into the books. Select a sample of shipping documents from the days just before year-end and trace each one forward to the sales journal and A/R ledger. Every shipment should have a corresponding invoice and receivable. A shipment that was never billed means both revenue and receivables are understated.

Check for Pledged, Factored, and Related Party Receivables

Receivables that look like ordinary assets on the balance sheet may actually be encumbered. Companies sometimes pledge receivables as collateral for a loan, or sell them outright to a factoring company for immediate cash. If either has happened without proper disclosure, the financial statements mislead about the assets available to general creditors.

Ask management directly whether any receivables have been sold, assigned, or pledged during the period. Review loan agreements and credit facility documents for language granting lenders a security interest. A sharp, unexplained decline in receivables relative to the revenue level is worth investigating; it can signal that receivables are being sold off the books.

If a factoring arrangement exists, determine whether the sale was with recourse or without. In a recourse arrangement, the company retains the risk of nonpayment and may need to record a liability even after removing the receivable from the balance sheet. Verify the relationship between the company and the factor, since factoring between related parties raises additional concerns about the economics and arm’s-length nature of the arrangement.⁴

Any pledged or factored receivables must be disclosed in the footnotes. Verify the disclosure covers the nature of the arrangement, the carrying amount involved, and any obligations the company retains.

Receivables owed by related parties get extra skepticism. A balance owed by a subsidiary, an officer’s family member, or a company under common ownership carries a different collection profile than one from an independent customer, and these balances create room for manipulating revenue or parking fictitious sales.

For each material related party receivable, read the underlying agreement and evaluate whether the terms make business sense. Check whether the transaction was authorized under the company’s policies for related party dealings and whether any exceptions were granted. If the financial statements claim the transaction was on arm’s-length terms, gather evidence for or against that claim by comparing the pricing, payment terms, and credit conditions to what the company offers unrelated customers. If the evidence doesn’t support the assertion and management won’t modify the disclosure, a qualified or adverse opinion may be warranted.⁵

Pay attention to the related party’s actual ability to pay. A large receivable from a related entity with no revenue and no assets is essentially worthless, whatever the contract says.

Review Subsequent Events

The audit doesn’t stop at the balance sheet date. Events between year-end and the report date can provide critical evidence about conditions that already existed at year-end. A customer’s January bankruptcy filing almost certainly reflects financial deterioration that was underway in December, and that kind of event requires the company to adjust the financial statements.⁶

The distinction that matters is whether the event reveals a pre-existing condition or creates a new one. A customer whose financial health was already declining before year-end and then files for bankruptcy afterward triggers an adjustment. A customer destroyed by a fire in February does not, because the fire didn’t reflect conditions at the balance sheet date. The company should disclose the fire but shouldn’t change the December 31 receivable balance.⁶

During the review, scan for major customer defaults, significant credit memo activity, unusual return volumes, and any new information about disputes that were pending at year-end. All of it feeds back into the allowance evaluation.

Watch for the Common Fraud Patterns

Accounts receivable is one of the most common targets for financial statement fraud, because inflating receivables inflates revenue at the same time. A few schemes recur often enough that they deserve dedicated attention.

Fictitious Sales and Channel Stuffing

The most direct form is booking sales to customers who never ordered anything, or recording revenue for goods that never shipped. Channel stuffing is a subtler variant: the company ships far more product to distributors than end-users will buy, often sweetened with deep discounts, extended payment terms, or side agreements allowing returns. Revenue looks real in the current quarter and reverses later through returns and write-offs.

Watch for receivable balances growing faster than revenue, days sales outstanding creeping up without a change in credit terms, and sales spikes in the last days of a period followed by heavy returns in the next. Direct confirmation with customers is the strongest defense against fictitious receivables, because a customer who never placed an order will flag the discrepancy.

Lapping

Lapping is an employee-level scheme in which someone steals a customer’s payment and covers the shortage by applying the next customer’s payment to the first account. The cycle continues and grows more complex over time. Signs include frequent reapplication of payments between accounts, persistent timing gaps between when cash arrives and when it’s posted, and customer complaints about incorrect balances or unexpected collection notices despite having paid.

Segregation of duties is the strongest preventive control, and confirmations often expose lapping because the customer’s records won’t match the company’s. If you suspect it, compare lockbox reports and bank deposits to A/R postings and look for patterns of manual intervention.

Delayed Write-Offs

Fraud sometimes runs the other way. A company delays writing off receivables it knows are uncollectible to avoid the bad debt expense, keeping reported earnings artificially high. Review old balances on the aging and ask management to explain why receivables more than 120 days past due haven’t been written off or specifically reserved. The explanation should be supported by evidence of ongoing collection activity or a genuine dispute.

Remember the GAAP-to-Tax Difference on Write-Offs

Under GAAP, credit losses are estimated in advance through the CECL allowance. The IRS does not allow this approach. For tax purposes, a bad debt deduction is only permitted when a specific debt actually becomes worthless.⁷

To claim a business bad debt deduction, the company must show the debt is uncollectible and that legal action would not produce payment. Factors supporting worthlessness include insolvency, a bankruptcy filing, disappearance, lack of assets, or repeated refusal to respond to collection efforts. A partial deduction is allowed for a debt recoverable only in part, provided the company charges off the appropriate amount on its books.⁸

This creates a temporary timing difference in the deferred tax accounts. When you audit receivables, verify the tax provision handles it properly. The GAAP allowance will almost always exceed the cumulative tax deductions taken, because GAAP recognizes losses earlier. If the deferred tax asset tied to the allowance looks out of proportion to the underlying timing difference, dig into the calculation.

1
Public Company Accounting Oversight Board. AS 2310 – The Auditors Use of Confirmation
2
Financial Accounting Standards Board. ASU 2025-05 Financial Instruments Credit Losses Topic 326
3
Financial Accounting Standards Board. Revenue Recognition
4
Internal Revenue Service. Factoring of Receivables Audit Technique Guide
5
Public Company Accounting Oversight Board. AS 2410 – Related Parties
6
Public Company Accounting Oversight Board. AS 2801 – Subsequent Events
7
Office of the Law Revision Counsel. 26 USC 166 – Bad Debts
8
Internal Revenue Service. Topic No 453 – Bad Debt Deduction