Existence Assertion

Where existence testing goes wrong

In 2023 the FRC published inspection findings where engagement teams claimed to have tested existence on payables by sending confirmation letters to suppliers. The supplier confirms a balance. The team ticks the box. But supplier confirmation on a payable tests accuracy, not existence. The real risk on payables is completeness (unrecorded liabilities), and the procedure ran backwards. The entire assertion was left untested while the file showed a green tick.

That finding keeps recurring because existence looks deceptively simple. "Does the item exist?" feels like a yes-or-no question. In practice, I think teams get it wrong more often than any other assertion, and the reason is directional confusion. They run a procedure, get a response, and map it to the wrong assertion in the workpaper.

The existence assertion is one of the assertions about account balances at period end defined in ISA 315 .A190(a). When management presents a balance sheet showing inventory of EUR 5 million, they are implicitly claiming that inventory is physically present or legally held at the reporting date. The auditor's job under ISA 500 .A31 is to test that claim in one specific direction: start from the recorded item, trace outward to independent evidence. Ledger to real world. Always.

What the standards require

ISA 315 .A190(a) lists existence as an assertion about account balances at period end. ISA 500 .A31 establishes the directional logic: to test existence you select from the financial statements and trace to evidence. To test completeness you go the other way, from an external source into the ledger. The two are mirror images, and the standard makes this explicit because confusing them is so common.

For inventory, ISA 501 .4 requires attendance at physical counting. ISA 501.8 covers inventory held by a third party, where physical inspection is not possible and the auditor must obtain confirmation or perform alternative procedures. ISA 505 governs external confirmations for receivables, bank balances, and other balances where third-party evidence is the most direct way to test existence.

None of this is optional. But the standards describe what to test, not how hard to push when the evidence is ambiguous. That gap is where most of the real problems live.

What actually happens on engagements

On a typical mid-market audit, existence testing runs on autopilot. The team pulls last year's file, updates the dates, sends the same confirmation letters to the same counterparties, and calls it done. I have reviewed files where the confirmation sample was identical three years running. No one asked whether the population had changed, whether new counterparties had appeared, or whether the response rate (which had dropped from 80% to 40%) meant the procedure was losing its power. That is a tick box exercise, not an audit.

Receivables confirmations are a good example. The team sends 30 letters. Twelve come back agreed. Eight come back with differences that get "resolved" by looking at subsequent cash receipts (which tests valuation and cut-off, not existence). Ten get no response, so the team performs "alternative procedures" that amount to checking subsequent cash. The workpaper concludes existence is tested. But for those ten non-responses, subsequent cash only tells you money arrived. It does not tell you the receivable was real at year-end. The customer could have paid against a different invoice entirely.

Inventory is worse. Teams attend the count, tick off their sample items, photograph the warehouse, and leave. Six months later they are documenting the work and cannot explain why the count sheets they observed do not reconcile to the final inventory listing. The reason is usually that management made post-count adjustments, and the team never traced those adjustments back. ISA 501.4 requires the auditor to evaluate management's final inventory records, not just the count. Most files I have seen stop at the count.

The grey zone: when existence gets hard

Existence testing is straightforward when you can touch the asset or get a letter back from a bank. It becomes genuinely difficult with line items where "existence" is a legal or contractual concept rather than a physical one.

Consider a provision for a legal claim. Management says a liability of EUR 2 million exists for an ongoing lawsuit. What does existence mean here? You cannot inspect the provision. You can read the legal correspondence, confirm the case is active, and review counsel's assessment. But the provision is an estimate of an uncertain outcome. Existence and valuation blur together in a way the assertion framework does not handle cleanly, and I think the standards underserve this area because they were written with tangible assets in mind.

Intangible assets create a similar problem. A capitalised development cost "exists" if the project meets the recognition criteria in IAS 38. But proving existence means proving the project is still viable, which is really a valuation question dressed up as existence. Teams that treat this as a simple tick are missing the point. The right approach is to acknowledge that for certain line items, existence cannot be tested in isolation from valuation, and to design procedures that address both simultaneously.

There is a legitimate disagreement among practitioners about how far the auditor should go with third-party-held assets. ISA 501.8 says obtain confirmation or perform alternative procedures. But what counts as an adequate alternative when the third party is in a jurisdiction where confirmations are unreliable? Some firms accept a warehouse receipt plus a contract review. Others insist on a site visit or an independent stock report. The standard leaves room for judgment, which means two reasonable auditors can reach different conclusions about whether existence has been adequately tested on the same balance.

A worked example: inventory at three locations

A distributor reports EUR 12 million of inventory across three locations: a company-owned warehouse, a third-party logistics (3PL) provider, and goods in transit from a supplier in Asia.

For the company warehouse (EUR 8 million), the team attends the year-end count, selects 40 items from the count sheets, and traces them to physical stock. All 40 items are found. Good so far. But during the count, the team notices that two aisles are roped off as "damaged goods pending write-off." Management says those items are excluded from the listing. The team accepts this without verifying it. Three months later, during the file review, you discover those damaged goods are still on the inventory listing at full value. The count tested existence for the items sampled but missed a EUR 300,000 overstatement sitting in plain sight.

For the 3PL location (EUR 3 million), physical attendance is not practical. The team sends a confirmation to the 3PL provider, who responds confirming "approximately 3 million EUR of product on hand." That word "approximately" should trigger follow-up. What does the 3PL actually hold? Does their record match the client's perpetual inventory by SKU? The team that accepts a round-number confirmation without reconciling it to the detail has not tested existence in any meaningful way.

For goods in transit (EUR 1 million), existence depends on when title transfers. The team checks the shipping terms (CIF destination) and confirms the goods have not yet arrived. Under CIF destination, title has not transferred. Those goods should not be on the client's balance sheet at all. The existence assertion fails not because the goods are fictitious but because they belong to someone else at the reporting date. This is where existence and rights and obligations overlap, and teams that test existence in isolation will miss it.

Second-order effects of weak existence testing

When existence testing is superficial, the downstream effects compound. If an asset does not exist, every other assertion mapped to it is automatically wrong. The valuation is wrong because you are valuing something that is not there. The classification is wrong. The disclosure is wrong. Existence is the foundation assertion: if it fails, nothing above it holds.

This is why regulators treat existence failures differently from, say, classification errors. A misclassified asset is still an asset. A non-existent asset is a misstatement that calls into question management's integrity. Once you find one fictitious receivable, the entire receivables balance is contaminated, and the sample you drew is no longer sufficient. You need to expand, re-evaluate fraud risk, and potentially reconsider the entire engagement under ISA 240 . The cost of getting existence wrong is not proportional. It is exponential.

Key standard references

• ISA 315 .A190(a) lists existence as an assertion about account balances at period end.
• ISA 501 .4-8 covers attendance at physical inventory counting and the specific requirements for inventory held by third parties.
• ISA 505 governs external confirmations as evidence for existence of receivables, bank balances, and other balances.
• ISA 500 .A31 establishes the directional relationship between the direction of a test and the assertion it addresses.
• ISA 240 covers the auditor's responsibilities relating to fraud, relevant when existence failures raise questions about fictitious assets.

Related terms

Related tools

Related reading

Frequently asked questions