November 4, 2026
9 min read
Credit Underwriting for MSME Loans: Process, Models, and the Data Behind Each Decision
November 4, 2026
9 min read
Two lenders can assess the same MSME and reach opposite conclusions. The difference is rarely the borrower. It is the underwriting: which data each lender trusts, which rules it applies, and where it lets an analyst override the system.
Credit underwriting is the decision engine of a lending business. In MSME lending, it has to work with thin bureau files and uneven financial statements, alongside a growing supply of transaction data from GST returns and the Account Aggregator framework.
This guide explains credit underwriting, the process lenders follow, the four underwriting models used in India, the data behind each decision, and how to design rules that hold up as volumes grow.
Credit underwriting is the process by which a lender assesses a borrower’s risk and decides whether to lend, how much, at what price, and on what conditions.
Every underwriting decision ends in one of four outcomes: approve, approve with conditions, refer to a higher authority, or decline. An approval also fixes the amount, tenor, pricing, security, and covenants.
Banks usually call this work credit appraisal, while NBFCs and fintech lenders more often say underwriting. The substance is the same. What differs is how much of the decision a system makes and how much an analyst makes.
Most MSME underwriting runs in five stages.
The last stage is the one most often skipped. Underwriting that never learns from its own portfolio repeats the same mistakes at larger volumes.
| Model | How it decides | Best suited to | Main weakness |
|---|---|---|---|
| Judgemental | An analyst reviews the file against policy and writes a memo | Larger tickets and complex structures | Slow and inconsistent across analysts |
| Scorecard | A statistical score from bureau and application data, with cut-offs by band | High-volume, small-ticket loans | Weak on thin-file MSMEs, and needs past performance data to build |
| Rule-based automated | Policy rules coded into a decision engine, with straight-through approval within limits | Standardised products and digital journeys | Rules go stale, and edge cases still need referral |
| Cash-flow underwriting | Repayment capacity read from bank and GST transaction data rather than collateral or bureau history | New-to-credit and collateral-free MSMEs | Only as good as the completeness and integrity of the data |
Few lenders use one model alone. A common design runs automated rules for screening and small loans, applies cash-flow analysis to every MSME file, and sends larger or flagged cases to an analyst.
Cash-flow underwriting matters most for India’s smaller businesses. Many have short or no commercial bureau histories and limited audited financials, but many file GST returns and run their business through a current account. Their transactions say more about repayment capacity than their balance sheets do.
| Data source | What it shows | Common pitfall |
|---|---|---|
| Credit bureau reports | Repayment history and current exposure of the entity and promoters, reported fortnightly since 1 January 2025 | Thin or missing files, and informal borrowing never appears |
| Bank statements | Cash flow, existing obligations and payment conduct | Partial accounts and edited PDFs |
| GST returns | Declared turnover, filing regularity and the spread of customers and suppliers | Non-GST income is missing, and filing gaps distort trends |
| ITR and financial statements | Profitability, net worth and leverage | Annual, lagged and self-reported |
| Udyam, MCA and KYB records | Identity, vintage, MSME category and directors | Static, and rarely updated |
The strongest decisions come from reconciling these sources, not from any one of them. Take an MSME with GST turnover of ₹4.6 crore, bank credits of ₹4.3 crore, and ITR turnover of ₹4.4 crore. The three agree within about 7%, so the underwriter can rely on turnover of roughly ₹4.4 crore. If the ITR showed ₹2 crore, the gap itself would be the finding.
For the bank statement side in detail, including which credits to strip out before any ratio, see our guide to bank statement analysis for lending.
Rules that work at 50 files a month often fail at 5,000. A few design choices keep them reliable:
An illustrative rule set for a small-ticket MSME working capital product (thresholds vary by lender):
| Rule type | Example | Action when triggered |
|---|---|---|
| Knock-out | Business vintage under two years | Decline |
| Knock-out | Statement fails integrity checks and cannot be refetched through Account Aggregator | Decline or refer |
| Scoring factor | Gap between bank credits and GST turnover | Score falls as the gap widens |
| Referral trigger | More than two inward returns in six months | Analyst review |
| Limit rule | Working capital capped at 20% of assessed turnover for MSE limits up to ₹5 crore | Limit set at the lower of the rule and the request |
For larger working capital limits, the MPBF method replaces the turnover rule in the last row.
RBI leaves underwriting design to lenders, but sets firm boundaries around it.
FinEye provides the data layer for cash-flow underwriting. Its Bank Statement Analyser and GST Analyser turn statements and returns into categorised data and ratios, delivered in PDF and JSON for a decision engine or loan origination system. The Account Aggregator module brings statements in with the borrower’s consent.
Underwriting quality is decided less by the model a lender picks than by the data it trusts and the discipline it applies to its own rules. A simple rule set on verified data usually beats a sophisticated model on documents nobody checked.
For MSME lending, that points one way. Bank and GST transactions, fetched with consent and reconciled with bureau and tax data, give underwriters a view of repayment capacity that collateral never could. Lenders that build their rules around that evidence, and keep testing those rules against outcomes, can grow volumes without growing losses at the same pace.
To add cash-flow evidence to your underwriting, see FinEye in action.
Frequently Asked Questions
Credit underwriting is the process a lender uses to assess a borrower’s risk and decide whether to lend, how much, at what price and on what conditions. It draws on bureau reports, bank statements, GST returns, financial statements, and the lender’s credit policy.
A typical process has five stages: pre-screening against policy, collecting and verifying data, analysis and scoring, the credit decision with limit and pricing, and post-disbursement monitoring that feeds back into the rules.
In practice, they describe the same work. Banks in India usually say credit appraisal for the full assessment of a loan proposal, while NBFCs and fintech lenders more often say underwriting, especially for automated or digital decisions.
Cash-flow underwriting judges repayment capacity from the borrower’s actual inflows and outflows, mainly bank statements and GST data, rather than relying on collateral or a long bureau history. It suits new-to-credit and collateral-free MSME loans.
Standardised small-ticket products can be approved straight through within coded rules. Larger or unusual files, and any file that triggers an integrity or mismatch flag, still need analyst review. Under RBI’s rules, the lender stays responsible for the decision either way.