GST Reconciliation for Credit Underwriting: GSTR-1 vs 3B vs 2B Red Flags Lenders Must Check

For most Indian MSMEs, GST returns are the closest thing to a monthly, third-party-anchored record of business activity. But a single GST number on a loan application, such as "annual turnover Rs 4 crore", says very little on its own. The value for a credit underwriter lies in reconciling the returns against each other and against the bank account. That is where inflated revenue, fake invoicing and stressed businesses show up.

This guide explains what each GST return contains, how lenders reconcile GSTR-1, GSTR-3B and GSTR-2A/2B, where e-way bills fit, how to compare GST turnover with bank credits, and the red flags that should send a file back for deeper review. For the broader picture of GST-based scoring, see our guide to AI credit scoring with GST data.

Why GST Analysis for Loans Needs Reconciliation, Not Just Extraction

Many lenders already pull GST data, through borrower-consented GSP or Account Aggregator routes, or from returns uploaded by the applicant. The common mistake is to stop at extraction: read the turnover from GSTR-3B, check that returns were filed, and move on. That approach misses the most useful signal GST offers, which is cross-consistency.

GSTR-1 is reported by the borrower, GSTR-2A and 2B are populated from what the borrower's suppliers report, e-way bills are generated at the time goods move, and bank statements are produced by the bank. A business inflating its revenue for a loan can manipulate one of these sources relatively easily, but keeping all of them aligned across 12 to 24 months is much harder. Reconciliation turns GST from a declared number into evidence.

What Each GST Return Tells a Lender

The Core Reconciliations Every Underwriter Should Run

1. GSTR-1 vs GSTR-3B: reported sales vs declared liability

Taxable value reported in GSTR-1 should broadly match the outward taxable value declared in GSTR-3B for the same period, after adjusting for amendments and credit notes. Small timing differences are normal. Persistent gaps, especially where GSTR-1 is materially higher than GSTR-3B, suggest the business is reporting invoices to support its buyers' ITC while under-declaring its own liability, which is both a compliance risk and a sign of cash stress. The reverse pattern, GSTR-3B consistently higher than GSTR-1, can indicate B2C sales not being invoiced in detail, or errors that will need correction.

2. GSTR-2B vs ITC claimed in GSTR-3B

Under the current framework, ITC claimed in GSTR-3B is expected to align with what is available in GSTR-2B, subject to prescribed conditions. Claims consistently above the 2B amount suggest the business is relying on invoices its suppliers have not reported, a classic marker of fake or unverified purchase invoices. Underwriters should compute the ratio of ITC claimed to ITC available month by month and investigate sustained excess.

3. GSTR-2A/2B purchases vs GSTR-1 sales

For a trading or manufacturing business, purchases and sales should move together, with a margin that fits the industry and HSN mix. Sales growing sharply while purchases stay flat, or purchases far exceeding sales for several months without a visible build-up of inventory, both warrant explanation. Comparing HSN codes on the purchase and sales side also reveals whether the business is actually trading in what it claims.

4. E-way bills vs invoiced goods sales

For goods businesses, e-way bill values for outward movement should be consistent with the B2B goods sales reported in GSTR-1, allowing for supplies below the threshold and intra-city exemptions in some states. Large invoiced sales with very few e-way bills, or e-way bills generated and then cancelled in volume, can indicate invoices without physical supply.

5. GST turnover vs bank credits

This is the reconciliation lenders rely on most. Gross receipts implied by GST turnover (including tax) should broadly reconcile with business credits in the borrower's bank accounts, after removing loan disbursals, inter-account transfers, refunds and capital infusions. GST challan debits in the bank statement should match the cash tax paid shown in GSTR-3B. Our bank statement analysis guide covers how to clean and classify credits before running this comparison; see also the bank statement analysis glossary entry.

GST Red Flags in MSME Underwriting

None of these flags is automatically disqualifying. Each needs context: industry, business model, seasonality and the borrower's explanation. But clusters of flags across returns are a strong signal for deeper review, a field visit or a reduced exposure.

The identity flags are where GST analysis and field verification reinforce each other. A business CPV that finds no stock, signage or activity at the GST principal place of business is a much stronger signal when combined with ITC anomalies. Our CPV guide and CPV verification page explain how business verification is structured, and our CPV TAT benchmarks show how to keep that step from slowing the file.

From Reconciliation to Credit Decision

GST reconciliation does not replace financial analysis; it calibrates it. In GST based underwriting, lenders typically use the reconciled figures in three ways:

• Assessed turnover: Use the lower of reconciled GST turnover and cleaned bank credits, or a policy-defined blend, as the base for eligibility instead of the applicant's declared figure
• Margin and cash flow: Apply an industry- or HSN-appropriate margin to assessed turnover, then test repayment capacity through FOIR for proprietors and DSCR for businesses
• Risk grading and conditions: Use red flag counts and severity to adjust the risk grade, cap the loan amount, require additional security or trigger enhanced CPV

The findings should be summarised in a GST section of the credit appraisal memo: months analysed, filing compliance, month-wise GSTR-1 vs 3B vs bank credit variance, ITC claimed vs 2B, top customers and suppliers, and each red flag with the underwriter's resolution. This creates an audit trail that internal audit, the credit committee and, in co-lending arrangements, the partner bank can rely on. See our credit underwriting process guide for how this fits into the full appraisal.

Practical Tips for Running GST Reconciliation at Scale

• Analyse at least 12 months, ideally 24: Shorter windows hide seasonality and pre-application window dressing
• Pull data with explicit consent: Use borrower-consented routes and record purpose and retention, consistent with the RBI Digital Lending Directions and data protection obligations
• Cover all GSTINs under the PAN: Businesses operating in several states have separate registrations; analysing one can understate or overstate activity
• Define tolerance bands in policy: Set acceptable variance ranges between returns and between GST and bank credits by segment, so analysts apply them consistently
• Automate the arithmetic, not the judgement: Let the system compute variances and flags; keep the explanation and override with a named underwriter
• Monitor after disbursal: Falling GSTR-3B turnover, late filing or ITC anomalies post-disbursal are useful early warning signals

GST Reconciliation in Co-Lending

In co-lending partnerships, the bank typically expects to see the evidence behind the originator's assessment rather than just the outcome. A structured GST reconciliation output, with month-wise variances and documented flag resolutions, is far easier for a partner to review independently than an analyst's free-text note. We discuss what partner banks typically expect in co-lending underwriting in India.

CARMA Assess runs GSTR-1, GSTR-3B, GSTR-2A/2B and bank statement reconciliation automatically as part of the credit underwriting workflow, and passes the reconciled figures and flags into the CAM, while CARMA Core carries them into the loan record for ongoing monitoring.