What financial statement fraud looks like in commercial underwriting
Financial statement fraud is the intentional misstatement, omission, or manipulation of borrower financials to influence a credit decision. In commercial lending, that can mean inflated revenue, understated liabilities, premature revenue recognition, hidden related-party activity, or expenses moved off the income statement to improve EBITDA.
Fraud is different from error. A borrower may have a bookkeeping issue, a timing difference, or an accountant reclassification that explains the mismatch. Your job in credit analysis is to test whether the inconsistency is explainable, repeatable, and supported by source documents.
The motives are familiar: meet use covenants, qualify for a larger facility, mask cash-flow stress, or avoid a personal guarantee. A borrower showing 18% revenue growth on internal statements while tax returns show flat sales and bank deposits are down 9% has a triangulation gap that should trigger enhanced diligence.
Start with financial spreading before judging the story
Before you decide whether a borrower’s story makes sense, normalize the data. Income statements, balance sheets, cash-flow information, and tax return schedules need to sit in consistent periods and categories. That is what financial spreading gives the underwriter: a stable base for comparison.
This matters because borrowers submit PDFs, Excel workbooks, scans, accountant-prepared statements, and tax packages that rarely line up perfectly. Without normalized spreading, a rent expense reclassification may look like margin improvement. With consistent mapping, EBITDA add-backs, operating expense shifts, and working-capital changes become visible across periods.
Compare interim statements, year-end financials, and prior-year submissions. Look for reclassifications that change EBITDA, current assets, use, or covenant compliance. Document every adjustment so the credit file can explain why a discrepancy is immaterial, operational, or suspicious.
| Unnormalized review | Normalized spreading | |
|---|---|---|
| Period comparison | Interim, year-end, and tax periods are hard to compare. | Periods are mapped consistently so trend breaks stand out. |
| Expense shifts | Reclassifications can appear as real margin improvement. | Expense movement is visible by category and period. |
| EBITDA review | Add-backs may be accepted without context. | Add-backs are tied to source documents and prior treatment. |
| Credit file support | Analyst judgment is harder to reconstruct later. | Adjustments are documented for review and approval. |
Revenue red flags lenders should test first
Revenue is often the first place to look when you want to detect financial statement fraud. Test whether reported sales growth is supported by bank deposits, receivables aging, sales tax filings, customer concentration, and industry demand. If the borrower says demand is rising, the operating evidence should move in the same direction.
Quarter-end or year-end spikes deserve extra attention, especially when they come with round-dollar invoices, manual journal entries, or rising days sales outstanding. If sales rise 25% while cash deposits rise only 3% and DSO moves from 42 to 78 days, underwriting should test collectability and possible premature revenue recognition.
Gross margin can also reveal the problem. If revenue grows and margin expands while payroll, input costs, inventory levels, or shipping activity do not support higher volume, ask for customer-level detail. Contracts, milestone billing, rebates, and related-party sales all require closer review because timing and collectability can change the credit picture.
Balance sheet signals that reveal hidden stress
The balance sheet often shows stress before the income statement admits it. Inventory growth that outpaces sales is a common warning sign, especially when borrowing base availability depends on inventory or accounts receivable. A borrower with inventory up 40%, revenue down 6%, and gross margin improvement is showing an unlikely combination that needs explanation.
Receivables quality should be tested through aging, subsequent collections, customer concentration, credits, and write-offs after period end. A growing receivables balance may be normal in a seasonal business, but it is a different issue when old balances are not converting to cash or when one customer accounts for a large share of sales.
Watch for capitalized expenses, unexplained intangible assets, related-party receivables, and sudden changes in accrued liabilities. Compare debt schedules to bank statements and UCC searches so you can identify undisclosed debt, lender payments, merchant cash advances, or payment deferrals that do not appear in borrower-prepared schedules.
Cash-flow and DSCR checks that expose manipulated earnings
Earnings manipulation becomes harder to defend when cash flow does not follow. Reconcile EBITDA to operating cash flow and pay close attention to sustained EBITDA improvement without cash conversion. A profitable borrower that needs constant revolver support, delayed vendor payments, or tax arrears may be weaker than the income statement suggests.
Stress-test DSCR using verified cash collections, recurring expenses, and actual scheduled debt service instead of borrower-adjusted earnings alone. A borrower reporting 1.45x DSCR on adjusted EBITDA may fall below 1.00x when non-recurring add-backs are removed and actual debt service is included.
Scrutinize add-backs for owner compensation, one-time expenses, management fees, and related-party charges. Some are valid, but they should be supported by history, policy, and source documents. Compare the final repayment capacity view against covenant compliance, overdrafts, tax arrears, and line use to see whether liquidity pressure is already visible.
Figures are illustrative and show how repayment capacity can change when unsupported add-backs and actual scheduled debt service are included.
Document-level fraud indicators in statements, tax returns, and bank records
Document review is not just clerical work. Inspect metadata, formatting, fonts, page numbering, signatures, accountant letters, and audit opinions for alteration clues. Missing pages, partial periods, repeated scans, or inconsistent entity names are process red flags that warrant direct-source verification.
Cross-check tax returns against financial statements for revenue, depreciation, officer compensation, related-party payments, and retained earnings movement. Then use bank statements to validate deposits, NSF activity, loan payments, merchant processor flows, and circular transfers between related accounts.
An accountant-prepared statement may show clean year-end liquidity while bank statements reveal repeated overdrafts and same-day transfers from an affiliate. Automated document ingestion can help standardize messy statements, tax returns, and bank records so the underwriter spends less time sorting files and more time testing contradictions.
Build an underwriting fraud detection workflow that is fast and defensible
A good fraud detection workflow is tiered. Every borrower gets baseline spreading and standard ratio checks. Anomalies trigger enhanced diligence. Unresolved contradictions move to escalation under credit policy, with the file showing what was requested, what was received, and what remains unsupported.
Set red-flag thresholds by loan size, industry, collateral type, and borrower complexity. A receivables aging issue in a small term loan may call for a narrower review, while the same issue in an asset-based facility may affect collateral value and availability. Consistency matters because two analysts should not reach different diligence standards on the same facts.
AI can speed up this work when it is explainable and kept inside the lender’s control framework. Crediflow AI ingests financial statements, tax returns, bank statements, PDFs, Excel files, and scans, then standardizes the data, runs ratio, cash-flow, and DSCR analysis, supports due diligence, and generates lender-branded credit memos. For regulated lenders, the right model is AI alongside existing loan origination systems and human approval judgment, not AI replacing credit policy.
Crediflow AI can move lenders from messy documents to a credit decision in minutes, with a full credit assessment in under 10 minutes and up to a 90% reduction in time-to-decision. That speed matters most when it improves consistency: the underwriter can focus on the contradictions that change repayment capacity, collateral value, and approval risk.
Frequently asked questions
How do lenders detect financial statement fraud during underwriting?
Lenders detect financial statement fraud by triangulating borrower financial statements against tax returns, bank statements, debt schedules, receivables aging, and operational data. The goal is to identify contradictions, such as revenue growth without matching cash deposits or EBITDA improvement without cash conversion, then document whether each issue is explainable or requires escalation.
What are the most common financial statement fraud red flags?
Common red flags include sudden revenue spikes, rising receivables days, inventory growth that outpaces sales, aggressive EBITDA add-backs, unexplained related-party balances, and missing or altered source documents. In underwriting, the most serious warning signs are those that affect repayment capacity, collateral value, or covenant compliance.
Can financial spreading help identify fraud?
Yes. Financial spreading standardizes statements into consistent categories and periods, making it easier to spot unusual reclassifications, margin shifts, use changes, and cash-flow gaps. It does not prove fraud by itself, but it creates the baseline needed for reliable anomaly detection.
How do bank statements help verify borrower financials?
Bank statements help validate whether reported revenue turns into actual cash deposits, whether expenses and debt payments match disclosures, and whether there are overdrafts or circular transfers. They are especially useful when internal financial statements are unaudited or borrower-prepared.
What should an underwriter do when fraud indicators appear?
The underwriter should request direct-source documentation, reconcile the discrepancy, and escalate unresolved issues according to credit policy. Depending on severity, the lender may reduce exposure, require additional collateral or guarantees, reprice the risk, or decline the deal.
Can AI detect financial statement fraud in lending workflows?
AI can help by ingesting documents, standardizing financial data, running ratio and DSCR checks, and flagging inconsistencies across statements, tax returns, and bank records. For regulated lenders, the strongest approach is explainable AI alongside human underwriting judgment and existing approval controls.