GST Return Analysis for Lending: A Complete Framework for Indian Credit Teams

GST returns give a lender a monthly, third-party-timestamped record of a borrower's sales, purchases and tax discipline. GSTR-1 shows what was invoiced, GSTR-3B what tax was actually paid, and GSTR-2A/2B what suppliers reported against the borrower. Read together they reveal turnover, seasonality and liquidity stress.

Key facts

• The data set is now near-universal for formal-sector borrowers. "The number of GST taxpayers increased from 66.5 lakh in 2017 to 1.65 crore as on May 2026," against gross FY 2025-26 collections of "~₹22.27 lakh crore" (PIB, *Nine Years of GST*, 30 June 2026).
• YuSight has processed 1 Mn documents, including GST returns arriving as portal PDFs, JSON downloads and phone photographs of a filed acknowledgement — classified, mapped to the right GSTIN and the right borrower entity, and spread without re-keying.
• GSTR-1 and GSTR-3B are locked together by rule. Under Rule 59(6)(b) a taxpayer cannot file GSTR-1 if the previous period's GSTR-3B is unfiled, and under section 39(10) GSTR-3B cannot be filed if that period's GSTR-1 is unfiled (GST portal, *FAQs > Form GSTR-1*). A gap in one return is therefore a gap in the whole chain.
• The annual audit trail has a turnover trigger. CGST Rule 80(3) requires a registered person whose "aggregate turnover during a financial year exceeds five crore rupees" to file a self-certified reconciliation statement in FORM GSTR-9C alongside the annual return, both by 31 December following the financial year (CBIC, *Rule 80, CGST Rules*).
• Goods movement is separately recorded. E-way Bill Rule 138(1) applies to "every registered person who causes movement of goods of consignment value exceeding fifty thousand rupees" (E-Way Bill Rules, Chapter XVI).

What does each GST return actually prove?

Analysts routinely treat "GST data" as a single object. It is not. Each return is filed by a different party, at a different time, with a different incentive, and proves a different thing.

The practical rule: GSTR-1 is a claim, GSTR-3B is a payment, GSTR-2B is a counterparty's claim about the borrower. Only the last one is not written by the borrower.

Why does GSTR-1 versus GSTR-3B matter more than either alone?

Because the gap between them is a direct read on cash.

A borrower who declares ₹7.29 crore of outward supplies in GSTR-1 for October and ₹6.88 crore in GSTR-3B for the same month has told the government two different numbers about the same month's sales. The overwhelmingly common reason is not concealment of sales — the sales are already declared in GSTR-1, invoice by invoice, with the customer's GSTIN attached. The reason is that GSTR-3B is the return where tax gets paid, and the borrower did not have the cash.

That makes a persistent GSTR-1 > GSTR-3B pattern one of the earliest liquidity signals available to a lender, typically visible months before a bank statement shows cheque returns or a bureau report shows an SMA tag.

Three patterns and what each means:

• GSTR-1 > GSTR-3B, recurring, then squared up in GSTR-9 with a DRC-03 payment. Cash-flow deferral. The borrower is using the exchequer as a short-term lender. Quantify the interest cost and treat the unpaid tax as a senior, undisclosed creditor.
• GSTR-3B > GSTR-1, recurring. Usually a data-quality problem — invoices reported in the summary but not uploaded — or reverse-charge and advance-receipt entries. Less alarming, but it degrades every turnover number you derive from GSTR-1.
• Both filed, both matching, both late. Not a turnover problem at all. It is a controls problem, and it says something about the finance function you are lending behind.

What do GSTR-2A and 2B tell you about the borrower's supply chain?

These are the only GST returns a borrower cannot author. They are built from what the borrower's suppliers filed.

Use them for four things:

1. Supplier concentration. If 61% of input value comes from two GSTINs, the borrower's cost base is hostage to two relationships. Check whether either shares a PAN prefix, address or director with the borrower.
2. Related-party purchases. Match supplier GSTINs against the group structure disclosed in the credit appraisal memorandum. Purchases routed through a promoter-owned entity at a mark-up are a leakage the audited financials will not flag.
3. Input mix against claimed activity. A borrower claiming to be a garment manufacturer whose GSTR-2B is dominated by trading purchases of finished goods is a trader with a factory address.
4. Supplier distress. Suppliers whose invoices stop appearing, or appear months late, are themselves under filing stress. Concentrated supply from stressed vendors is a real operational risk.

Reconcile against 2B, not 2A. 2A keeps moving as suppliers file late, so a reconciliation done on 2A cannot be reproduced next week. 2B is frozen on the 14th.

How do you reconcile GST turnover with the ITR and audited financials?

This is the reconciliation credit committees ask for and analysts most often get wrong, because they compare two numbers that were never meant to be equal.

Take Vasant Agro Foods Pvt Ltd, a Nashik-based food processor, FY 2024-25. Single GSTIN in Maharashtra plus one in Telangana for a distribution branch.

Step 1 — the borrower's own three numbers

The gap is not spread evenly. It sits entirely in September, October and November 2024, and it was regularised in the GSTR-9 for the year with additional tax discharged through DRC-03. That is the liquidity story, and it belongs in the CAM's risk factors, not in a footnote.

Step 2 — bridge GST turnover to audited revenue

A residual of ₹10,00,000 on ₹50.08 crore is 0.2%. That is a clean reconciliation. What made it clean was not the borrower's honesty — it was doing the four bridging adjustments before comparing.

What does a turnover mismatch actually mean?

Not fraud, in most cases. Before anyone writes "revenue overstatement" into a note, work through the ordinary causes:

• Branch and stock transfers. Supplies between distinct persons under the same PAN are taxable events under GST but not revenue under Ind AS or AS 9. Multi-state borrowers will always show GST turnover above book revenue.
• Schedule III and non-GST items. Sale of land or a completed building, salary, and certain other transactions sit outside GST but may sit inside the ITR.
• Other income. Scrap sales, asset disposals, rental income and interest may be GST-taxable while being reported below the revenue line.
• Timing. GST follows time-of-supply rules; the P&L follows revenue recognition. March-versus-April cut-off differences appear every single year.
• Exempt and nil-rated supplies. Unbranded food grain, agricultural produce, education and healthcare services appear in GSTR-1 totals but not in the taxable-supply lines of GSTR-3B.
• Credit notes and rate-difference adjustments issued after year end but relating to the prior year.
• Composition-scheme or multiple GSTINs. One GSTIN's turnover is not the entity's turnover. Pull every GSTIN under the PAN.

The mismatches that genuinely warrant escalation are the ones that survive all of the above: a residual that is large, one-directional and repeated across years; a borrower who cannot explain the bridge; or an entity whose GST turnover falls while its audited revenue rises. The last one is the most serious, because inflating book revenue costs nothing in tax and inflating GST turnover costs 5-18% in cash.

How do you read seasonality from monthly filings?

Annual turnover hides the working capital cycle. Monthly GSTR-3B does not.

Vasant Agro Foods, FY 2024-25 outward supplies per GSTR-3B, in ₹ crore:

Read: October alone is 13.5% of the year's declared supplies (6.88 of 50.92), and October to December together are 32.8% (16.70 of 50.92). Peak despatch is festive plus kharif arrival. That means:

• Peak inventory build runs August-September, one to two months ahead of despatch, so the borrower's genuine working capital peak is not in March.
• A drawing power assessment based on a 31 March stock and book-debt statement will understate the facility this borrower actually needs in September, and a sanctioned limit sized on the March position will strand them in the busy season.
• Receivables peak in November-December. If the borrower's debtor days are 68, collections against October despatch land in December-January — which is exactly when a lender who has not read the seasonality will be worried about a quiet account.
• The March spike (4.89) is worth a question. A month-of-March jump above the trend is sometimes real year-end demand and sometimes invoicing pulled forward to hit a turnover covenant or a distributor scheme. Check it against March e-way bills.

Three years of monthly filings give you a seasonality index far more reliable than anything the borrower's projections in the CMA data will tell you, because the CMA projections are an argument and the returns are a filing.

Are filing delays a reliable stress signal?

They are one of the better ones, for a specific reason: filing GST returns is cheap and mandatory, so a borrower who stops is usually a borrower who cannot pay the tax rather than one who forgot.

What to build:

Note the shape. GSTR-1 stays roughly on time while GSTR-3B slips — the borrower keeps declaring invoices (which costs nothing and keeps customers' ITC flowing) but delays the return that requires cash. A widening GSTR-1-to-GSTR-3B filing gap is the same liquidity signal as the value gap, in a different dimension.

Escalating consequences to check for:

• Late fee and interest accruing as an undisclosed liability.
• ITC blocked for the borrower's customers, which makes the borrower a difficult vendor and can cost them the customer.
• Rule 59(6)(b) lockout — once GSTR-3B for a period is unfiled, the next GSTR-1 cannot be filed, so a single missed month cascades.
• E-way bill generation blocked for continued non-filing.
• Time bar on filing. Returns become non-filable three years after the due date under the Finance Act 2023 amendments, implemented on the portal from 2025.

What does a cancelled or suspended GSTIN mean for a live limit?

It means the borrower may not legally raise a tax invoice, cannot pass ITC to customers, and in practice will lose institutional buyers within a billing cycle.

Distinguish the three states:

• Suspended. Registration is on hold pending cancellation proceedings, often triggered by non-filing or by anomalies between the borrower's own returns. Business is restricted but the registration is not dead.
• Cancelled by the officer. Usually non-filing or a compliance failure. Revocation is possible within a window; check whether an application has been made.
• Cancelled voluntarily. The borrower closed the registration. On a live working capital limit, this needs an explanation and a document, not a phone call.

Practical checks: verify status on the GST portal's Search Taxpayer for every GSTIN under the PAN, not just the one on the application; check the date of cancellation against the period of the returns you have been given; and confirm the borrower's principal place of business on the registration matches the address in your KYC and the address on the hypothecated stock. A GSTIN registered at a residential address for a borrower claiming a 40,000 sq ft plant is a site-visit trigger.

How does e-way bill data fit in?

E-way bills record physical movement, which is the one thing invoices cannot fake cheaply. Rule 138(1) applies to consignments exceeding ₹50,000 in value.

Three uses:

1. Corroborate declared despatch. Aggregate e-way bill value by month against GSTR-1. Systematic excess of GSTR-1 over e-way bills, in a goods business, means invoices without movement.
2. Read the real activity calendar. E-way bills are generated at despatch; invoices can be dated to suit. Month-end and quarter-end invoicing spikes with no matching movement is the classic revenue-pull-forward pattern.
3. Map routes and counterparties. Ship-to locations and transporter IDs tell you where the goods actually go, which is useful when the customer list looks concentrated or unfamiliar.

Cancellation rates matter too. A high proportion of cancelled or expired e-way bills suggests either poor logistics discipline or invoices raised and reversed.

Can GST data replace audited financials?

No, and lenders who have tried have learned why.

GST returns tell you about turnover and tax discipline. They tell you nothing about cost structure, depreciation, interest cover, leverage, contingent liabilities, related-party balances, promoter drawings or net worth. You cannot compute DSCR or TOL/TNW from a GSTR-3B.

What GST data does exceptionally well is verify and time-stamp the top line that the audited financials assert, at monthly granularity, in a form the borrower cannot retrospectively edit. For small-ticket MSME lending where audited financials are provisional, stale or absent, GST-based surrogate underwriting is defensible — with tight ticket sizes and a hard requirement that the GST history be at least 12 and preferably 24 months long. For anything with a term structure, GST is a corroboration layer, not a substitute.

The strongest file uses all three: GST returns for turnover and rhythm, audited financials for structure, and Form 26AS for independent confirmation of who actually paid the borrower.

How YuSight handles GST returns

GST documents reach a credit file in every possible shape: portal PDFs, JSON exports, GSTR-9C in a scanned copy with a signature block, and screenshots. YuSight's Document Intelligence classifies each one, reads the GSTIN and the period off the face of the document, and maps it to the correct borrower entity — which matters when a group has eleven GSTINs across six states and the borrower has sent them in one folder with no naming convention.

From there, Financial Spreading pulls the monthly outward supply and ITC series, builds the GSTR-1-to-GSTR-3B comparison and the filing-delay calendar, and carries every figure into the CAM with a citation to the source page. Nothing is re-keyed, and nothing is asserted without a link back to the return it came from.

FAQ

How do lenders use GST returns in credit assessment?

Mainly three ways: to verify the turnover the borrower claims, to see the month-by-month rhythm of the business, and to spot cash stress through the gap between what was invoiced in GSTR-1 and what tax was actually paid in GSTR-3B.

Which GST returns matter most for underwriting?

GSTR-3B and GSTR-1, read side by side and month by month. GSTR-2B matters when you want to understand the supply chain, because it is built from what suppliers filed rather than from what the borrower says.

Can GST data replace audited financials?

For small-ticket, short-tenor MSME lending it can carry a lot of the load. For anything with a repayment schedule it cannot, because GST returns say nothing about costs, leverage, interest cover or net worth.

What does it mean when GSTR-1 is higher than GSTR-3B?

Usually that the borrower has declared the sales but has not paid the tax on them. The sales are already on record invoice by invoice, so this is a cash-flow signal rather than a concealment signal — and the unpaid tax is a senior creditor you should be counting.

Is a GST turnover mismatch with the ITR evidence of fraud?

Almost never on its own. Branch transfers, exempt supplies, other income, and March-versus-April timing differences produce mismatches in perfectly clean files. Do the bridge first; escalate only what survives it.

How do I check whether a borrower's GSTIN is active?

Use Search Taxpayer on the GST portal and check every GSTIN registered under the borrower's PAN, not just the one on the loan application. Look at the status and the effective date, and match the principal place of business against your own KYC address.

What is the difference between GSTR-2A and GSTR-2B?

GSTR-2A keeps changing as suppliers file late, so it is a moving target. GSTR-2B is frozen on the 14th of the following month, which makes it the one you should reconcile against if you want a result someone else can reproduce.

Do e-way bills tell a lender anything the returns do not?

Yes — whether the goods actually moved. Invoices can be dated to land in a convenient month; a truck either left the gate or it did not. Comparing monthly e-way bill value against GSTR-1 catches invoicing that has run ahead of despatch.

How far back should a credit team pull GST returns?

Twenty-four months is the working minimum, because you need two cycles to separate seasonality from deterioration. Twelve months tells you the shape of the year but not whether this year is worse than the last one.

Key takeaways

• Each GST return proves something different. GSTR-1 is a claim, GSTR-3B is a payment, GSTR-2B is what someone else said about the borrower.
• The GSTR-1 minus GSTR-3B gap, in both value and filing date, is one of the earliest liquidity signals a lender can see.
• Reconcile GST turnover to audited revenue through an explicit bridge — branch transfers, other income, exempt supplies, cut-off. Compare the residual, not the raw numbers.
• A mismatch is not fraud. A mismatch that is large, one-directional, repeated and unexplained is a different matter.
• Monthly filings give you a seasonality index that will reshape both the limit and the timing of the drawing power assessment.
• GST corroborates the top line. It does not replace the balance sheet.

Upload a messy document set and see it classified — eleven GSTINs, three years of returns, no naming convention. Book a live demo.