GST Turnover vs Bank Credits: The Variance Almost Everyone Calculates Wrong

Put a borrower's GST returns next to their bank statements and the numbers will not agree. This is often treated as a finding. It is not. The two figures measure different things over different periods on different bases, and they would disagree even if the borrower were perfectly honest.

The useful question is not whether there is a gap. It is what remains after you account for the reasons a gap must exist. That residual is the credit signal, and most lenders never calculate it — they compare gross to gross, get a large number, and either escalate a clean file or wave through a dirty one depending on which direction the error happened to point.

Here is how to construct the number properly.

Why the two figures can never match

Six structural reasons, before any question of misreporting arises.

GST is recognised on invoice date. Bank credits arrive on payment date. If the borrower sells on 60-day terms, this month's GSTR-1 describes invoices whose money arrives two months from now. On a growing business, banked credits will systematically trail declared turnover — permanently, and by more the faster it grows. Read as a same-month comparison, healthy growth looks like unbanked sales.

Declared taxable turnover excludes GST. The customer pays it. A business declaring ₹1 crore of taxable turnover at 18% issues invoices totalling ₹1.18 crore and, if it collects in full, banks ₹1.18 crore. Comparing bank credits against the taxable value line builds an 18% discrepancy into the baseline before you start. This single omission is the most common reason a variance calculation points the wrong way.

Not all revenue is taxable. Exempt supplies, nil-rated supplies and goods outside the GST net entirely generate banked receipts that never appear in taxable turnover.

Exports are zero-rated. They appear in GSTR-1 at nil tax, arrive in foreign currency, and frequently land in a separate account you may not have been given.

Marketplace settlements are net. An e-commerce seller declares gross sales value and receives that figure net of commission, TCS and returns. A structural 15–30% gap is normal and says nothing about the borrower.

Cash sales are declared but never banked. Standard in retail, food service and small trade.

Add to these the practical ones: the borrower has five accounts and gave you three, operates under multiple GSTINs across states, or has customers deducting TDS so that every receipt arrives net of the invoice value.

Building a real variance number

Five steps. The first is the one that determines whether any of the rest is meaningful.

1. Strip non-revenue credits from the bank side.

Gross credits are not sales. Loan disbursals, self-transfers between the borrower's own accounts, related-party inflows, refunds, interest, and investment redemptions all land in the same column as customer receipts.

This step is where the calculation usually fails, because it depends entirely on the categorisation underneath being correct. A loan disbursal classified as a trade credit does not merely inflate the bank side — it inverts the conclusion, because the borrower now appears to be banking more than they declare, which reads as GST under-reporting rather than as leverage. In our 200,001-transaction benchmark, loan disbursals misread as counterparty transfers were among the most frequent error classes we found.

2. Gross up the GST side. Apply the applicable rate to taxable turnover to arrive at expected billing, then add exempt, nil-rated and non-GST supplies, and net off credit notes.

3. Align the periods to the actual collection cycle. Estimate days sales outstanding from the observed lag between invoice-heavy months and credit-heavy months, then offset the two series by that lag. Comparing April GST to April banking is a category error unless the borrower sells for cash.

4. Subtract known structural leakage. Declared cash sales, marketplace netting, TDS deductions, and turnover under GSTINs or in accounts outside your view. Each should be documented, not assumed.

5. What remains is the residual. This is the only number worth escalating.

What this looks like with numbers

An SME declares ₹1.20 crore of taxable turnover for the month at 18% GST, sells on roughly 60-day terms, and has cash sales running at about 8% of billing.

• Expected billing: ₹1.20 cr × 1.18 = ₹1.42 cr
• Bank credits two months later, gross: ₹1.65 cr
• Less self-transfers ₹22 lakh, a loan disbursal ₹15 lakh, interest and refunds ₹3 lakh → net trade credits ₹1.25 cr
• Shortfall against expected billing: ₹17 lakh, or 12%
• Less documented cash sales of ~₹11 lakh
• Residual: ₹6 lakh, roughly 4%

Note what the naive comparison would have produced. Gross bank credits of ₹1.65 crore against declared turnover of ₹1.20 crore is 37% above declaration — a figure that reads as significant under-reporting to GST. The corrected answer is a 4% shortfall in the opposite direction, comfortably within tolerance.

The naive calculation was not merely imprecise. It pointed the wrong way.

Reading the residual

Once the residual is real, direction tells you what you are looking at.

Banked receipts exceed declared turnover. Either sales are being under-declared to GST, or money that is not sales is entering the account. The first creates a contingent tax liability that ranks ahead of you and is not on any balance sheet you have been shown. The second means the revenue you are lending against does not exist. Both are material; they call for different questions.

Declared turnover exceeds banked receipts. Sales are not reaching this account. Three possibilities: they are being collected in cash, they are landing in an account you have not seen, or — the credit-relevant one — they are not being collected at all.

That third case is where this analysis earns its keep, and it is invisible in either data source alone. A borrower with stable declared turnover and steadily falling collections is running a receivables failure. GST returns look fine because invoices are still being raised. The bank account looks merely quiet. The gap between them is the only place it shows.

The trend matters more than the level

A stable 15% variance is a business model. A variance widening from 8% to 22% over four months is a business breaking.

This has to be built as a monthly series and read month on month. An annual figure averages across the entire collection cycle and destroys exactly the signal you are looking for — a year in which collections deteriorated sharply from month seven onward can produce a perfectly ordinary annual variance. Year-on-year comparison is worse still, because it puts twelve months between the two observations and tells you nothing about direction of travel within the current cycle.

Three shapes worth watching in the monthly series:

• Widening gap, flat declared turnover. Collections deteriorating. This precedes visible distress by months.
• Narrowing gap, falling declared turnover. The business is shrinking but converting what it bills. Less alarming than it looks.
• Gap inverting from negative to positive. Non-trade money has started entering the account. Check what it is before you check anything else.

What each source sees that the other cannot

GST filings carry signals with no banking equivalent.

GSTR-1 against GSTR-3B. GSTR-1 reports invoice-level outward supplies; GSTR-3B is the summary on which tax is actually paid. Persistent excess of GSTR-1 over 3B means the borrower is invoicing without discharging the corresponding liability — an accruing statutory exposure, and a governance signal independent of anything in the bank account.

Filing discipline. Late filing frequency is a proxy for operational control. Nil returns filed by an allegedly operating business are a straightforward contradiction.

Seasonality. A monthly filing history establishes the borrower's real trading calendar — peak months, idle months, and the trajectory over cycles — which lets you judge whether a weak month is a pattern or a problem.

Banking carries what GST cannot: whether the money actually arrived, what left afterwards, what obligations were serviced, whether payments bounced, and what the balance looked like on the days that mattered. GST tells you what the borrower says they sold. Banking tells you what happened to the business. Neither is sufficient, and the residual between them is a third thing that neither contains.

Frequently asked questions

Should GST turnover match bank credits? No. GST is recognised on invoice date and excludes tax; bank credits arrive on payment date and include it. Add cash sales, exempt supplies, exports, marketplace netting, TDS deductions and accounts outside your view, and a substantial gap is normal. Only the residual after adjusting for these is meaningful.

How do you calculate a GST to bank variance correctly? Strip non-revenue credits from the bank side, gross up declared taxable turnover by the applicable GST rate and add exempt supplies, offset the two series by the borrower's actual collection cycle, subtract documented structural leakage such as cash sales, and read what remains.

What does it mean if bank credits exceed declared GST turnover? Either turnover is being under-declared, creating a tax liability ranking ahead of your exposure, or non-sales money is entering the account and being mistaken for revenue. Verify the categorisation of large credits before concluding under-declaration.

What does a widening GST to banking gap indicate? Most often a receivables collection failure. Invoices continue to be raised, so declared turnover holds steady, while the money stops arriving. Because the deterioration is visible in neither source alone, it typically appears in the variance series months before it surfaces in conduct or bureau data.

Can GST data be used for MSME credit assessment on its own? It establishes declared commercial activity, filing discipline and trading seasonality, but not liquidity, obligations or repayment conduct. Its value in underwriting comes from cross-verification against banking rather than from standalone use.

---

Fiscus reads GST filings and bank statements as a single borrower view, with non-revenue credits isolated and monthly variance tracked as a series. Book a parallel evaluation on cases your team has already assessed.