Revenue Recognition Fraud

Before 2014, revenue recognition rules under both GAAP and IFRS were scattered across dozens of industry-specific pronouncements. This fragmentation created legitimate diversity in practice and illegitimate cover for manipulation. The IASB and FASB replaced that patchwork with a single converged model: IFRS 15 (effective January 2018) and ASC 606 (effective for annual reporting periods beginning after 15 December 2017, i.e., calendar year 2018 for most public companies). The core is a five-step sequence that every revenue transaction must pass through.

1. Step 1: Identify the contract with the customer
   A contract must be approved, have commercial substance, and create enforceable rights. Side letters that modify these terms but are concealed from auditors undermine this step.
2. Step 2: Identify the performance obligations
   Distinct promises within the contract are separated. Bundling distinct services to obscure undelivered elements is a manipulation technique at this step.
3. Step 3: Determine the transaction price
   Variable consideration, discounts, and refund liabilities must be estimated and constrained. Ignoring side-agreement return rights inflates the transaction price.
4. Step 4: Allocate the transaction price
   Price is allocated to each performance obligation based on stand-alone selling prices. Reallocation toward delivered elements, as in the Xerox case, front-loads revenue.
5. Step 5: Recognise revenue when each obligation is satisfied
   Recognition occurs either at a point in time or over time as control transfers. Claiming control transferred before it genuinely did is the core of premature recognition.

Every revenue recognition scheme can be mapped to the step it corrupts. That mapping structures the investigation: find which step was bypassed, find the documentation that was created to support the bypass, and trace the cash flows that should have followed a legitimate transaction but did not.

Premature recognition is the most common form of top-line manipulation because the underlying sale is real: the customer exists, the invoice was raised, and cash will eventually arrive. The dispute is about timing, not existence, and management can often construct a superficially plausible accounting rationale.

Common premature recognition techniques include: recording a shipment as revenue before the customer acceptance period has expired; treating the full value of a multi-element contract as satisfied when only the hardware has been delivered; recognising revenue on long-term construction-style contracts ahead of the percentage-of-completion calculation; and booking consignment inventory as sold goods.

HealthSouth, the US rehabilitation services chain, ran a form of premature recognition from 1996 to 2002 that accumulated into an overstatement ultimately quantified at approximately $2.7 billion, though the SEC's initial complaint in March 2003 alleged at least $1.4 billion.

The scheme was run by the CFO and controller, who instructed accounts to make small fictitious or premature entries across thousands of accounts rather than large adjustments in a few. The dispersion made each individual entry look immaterial. The aggregate was catastrophic.

Fictitious revenue schemes require a fabricated customer, a fabricated invoice, and, to survive any serious scrutiny, fabricated cash movement. The cash-movement requirement is where many pure fabrication schemes collapse: an investigator who traces each large receivable to an actual bank receipt quickly identifies entries with no corresponding cash inflow.

Round-trip transactions solve the cash problem by routing funds through a related party or a controlled entity. Company A records a sale to Company B. Company B pays. Company A then separately transfers equivalent funds to Company B under a different label, completing the circle. Both companies record clean-looking receipts, but no genuine economic transaction occurred. Related-party registers and bank record analysis across connected entities are the standard investigation tools.

Bill-and-hold and channel stuffing both involve a physical transfer of goods that appears to satisfy a delivery condition while the genuine transfer of control and risk to the buyer has not occurred. The structural difference is location: bill-and-hold retains the goods at the seller's premises; channel stuffing ships them to a distributor who, because of concealed side agreements, functions as a temporary warehouse for the seller.

For bill-and-hold to be legitimate under IFRS 15 paragraph B81, the customer must have a substantive reason for the arrangement, the goods must be identified as belonging to the customer and ready for delivery, and risk of ownership must have genuinely passed. Fraudulent bill-and-hold arrangements fabricate the customer's written request without any of those conditions being genuinely met.

Channel stuffing is structurally harder to prove because the goods did ship. The investigation pivots to the side agreements, often called side letters, that modify the apparent sale: return authorisations, extended payment terms contingent on the distributor selling through, price protection guarantees, and marketing development fund payments that effectively reduce the net selling price retroactively. These documents are frequently not disclosed to auditors.

Xerox Corporation restated results in 2002 for the period 1997 to 2001, reducing pre-tax income by approximately $1.4 billion. The core scheme exploited the allocation step of long-term copier leases. Each lease bundled equipment, maintenance, and financing components.

Xerox inflated the portion of each contract's value allocated to the equipment, which was recognised upfront, and deflated the service component, which would have been recognised over the lease term. No external transaction was falsified. The manipulation was entirely in the internal allocation formula.

The SEC found that Xerox used four main techniques, collectively called the "topside" adjustments: geography swaps, finance income accelerations, margin normalisation, and ROE accounting adjustments. Each was applied by regional finance teams against a numerical target, and the targets cascaded from corporate headquarters. What looked like accounting judgment was a coordinated earnings management program run to a spreadsheet.

Revenue fraud leaves ratio signatures that an analyst can detect before obtaining a single internal document. These are not proof of fraud. They are priority indicators that justify a deeper investigation.

• DSO creep: accounts receivable growing faster than revenue. Genuine sales collect cash within their stated terms; fictitious or premature revenue accumulates as uncollected receivables.
• Revenue-to-cash divergence: operating cash flow consistently below net income. The accrual manipulation that inflates revenue does not create cash, so the two lines diverge.
• Quarter-end revenue spikes: a disproportionate share of annual revenue recorded in the final days of each quarter, inconsistent with the business's normal sales pattern.
• Rising return rates: for channel-stuffing schemes, abnormally high product returns in the period following a strong revenue quarter.
• Unusual journal entries: large round-number or period-end revenue entries posted by senior finance staff rather than through the normal order-to-cash cycle.

A single ratio flag in isolation is rarely diagnostic. A rising DSO in a business deliberately extending credit terms is benign. The same DSO movement alongside a widening revenue-to-cash gap and concentrated period-end journal entries, in a business with stable credit-term policy, is a materially different picture.