Content authored by practitioners with experience at Amazon India, Intuit QuickBooks, and the Tata Group. Meet the team →
Credit officers reading bank statements for risk manually miss NACH bounce patterns, round-trip transactions, and PDF authenticity issues — producing inconsistent underwriting decisions and documentation gaps that fail RBI inspection.
Follow a seven-step sequence: (1) verify PDF authenticity and balance arithmetic, (2) read opening/closing balance trend over the full statement period, (3) classify income streams excluding transfers and disbursals, (4) compute average monthly balance on 1st, 14th, and last day, (5) check NACH/EMI continuity and return codes, (6) scan narrations for 10 risk word categories, (7) compute FOIR against classified income with proposed EMI included.
12-month statement for MSME loans, 6-month for personal loans, 3-month for microfinance. FOIR threshold 50% retail / 55% MSME. NACH return threshold: zero returns in last 3 months. Balance check dates: 1st, 14th, last day of month.
A structured credit signal report covering income classification, FOIR, average monthly balance, NACH continuity status, round-trip flag, risk word category hits, and PDF authenticity verdict — with supporting evidence for each signal.
Reading a bank statement for credit risk starts before you look at a single transaction. The first question is whether the statement in front of you is authentic. The second is whether the balance trend tells a story that 30 seconds of scanning can reveal. The actual transaction-level analysis comes third.
What Reading for Credit Risk Means
Reading a bank statement for credit risk is structured signal extraction — not bookkeeping. The credit reader is not verifying that debits and credits balance. They are answering seven specific questions: Is this document authentic? Is the borrower’s financial position improving or declining? What income can be relied upon? What obligations already exist? Does the borrower manage their balance well? Are there stress indicators in the narrations? Can the proposed EMI be serviced?
Each question maps to a specific step in the review sequence.
The Seven-Step Credit Risk Reading Process
Step 1 — Verify PDF Authenticity
Before reading a single transaction, check that the closing balance at the end of each month equals the opening balance plus net transactions. Any month where this arithmetic does not hold indicates a potential edit. For Indian private bank PDFs (HDFC, ICICI, Axis), also check that font sizes are consistent across the narration and amount columns — edited cells often show a different font weight.
Step 2 — Read the Opening-to-Closing Balance Trend
Look at the closing balance on the last day of each month across the statement period. A declining trend over 6 to 12 months is a balance deterioration signal regardless of income level. A flat balance despite growing income suggests the borrower is consuming all cash. Rising balance with consistent income is the strongest positive signal.
Step 3 — Classify Income Streams
Separate all credit entries into: salary (regular, same-day monthly, single employer narration), business receipts (multiple counterparties, business narrations), rental income (recurring credits, fixed payors), and exclusions (inter-account transfers, loan disbursals, refunds). Average the classified credits across the statement period, removing outlier months.
Step 4 — Compute Average Monthly Balance on 1st, 14th, and Last Day
The average balance on these three dates per month reveals how the account is managed around the payment cycle. An account with a high monthly average but low balance on the 1st (the typical NACH debit date) will bounce mandates even if the average looks healthy. This three-date check takes 5 minutes manually; automated tools compute it across 12 months in seconds.
Step 5 — Check NACH and EMI Continuity
Identify all recurring debits that match NACH or ECS patterns — same amount, same originator, same date each month. Verify that each mandate executed without a return code across the last 6 months. NACH bounce codes to scan for in narrations: NACH-10, NACH-12, RTNACH, INS FND, or similar bank-specific abbreviations.
Step 6 — Scan Risk Word Categories
Scan narration entries for the 10 risk word categories: gambling platforms, alcohol distributors, crypto exchanges, informal lending, legal proceedings, round-trip counterparties, and others. A single hit is not an automatic reject — frequency and proportion of income determine the risk weight.
Step 7 — Compute FOIR
Total all identified recurring obligations and divide by average classified monthly income. Add the proposed new EMI to the obligations. Compare against the NBFC’s threshold (typically 50% for retail, 55% for MSME). Document the classified income figure and the obligations list — this is the FOIR evidence for the credit appraisal note.
Signal Reference Table
| Step | What You Are Checking | Red Flag Threshold |
|---|---|---|
| PDF authenticity | Balance arithmetic continuity | Any month where open + net ≠ close |
| Balance trend | Month-end closing balance over 12 months | Declining for 3 or more consecutive months |
| Income classification | Net classified monthly credit | Below 1.5x proposed EMI |
| Balance on debit dates | Avg balance on 1st, 14th, last day | Below proposed EMI amount on debit date |
| NACH continuity | Return codes in last 6 months | Any return in last 3 months |
| Risk word flags | Narration category hits | 3 or more hits in same category |
| FOIR | Obligations ÷ income (with proposed EMI) | Above 50% retail / 55% MSME |
India-Specific Considerations
NACH mandate execution and bounce patterns are India-specific signals that global credit risk frameworks do not address. The Financial Intelligence Unit India publishes suspicious transaction reporting guidance that informs the risk word category framework — particularly for gambling, crypto, and informal lending narration patterns.
PSU bank statements (SBI, Bank of Baroda, Union Bank) use abbreviated or non-standard NACH return codes that are not always obvious to analysts trained on private bank statements. Bank of Maharashtra uses “RTNACH” where HDFC uses “NACH-10” — the same underlying event labelled differently. Manual review trained on one bank’s format will misread another’s bounce signals.
A bank statement analysis platform that covers 34+ Indian banks normalises these narration differences across PSU, co-operative, and private bank formats — so the same NACH continuity check applies regardless of which bank issued the statement.
For lenders handling HDFC-heavy applicant flows, a bank statement analyzer India optimised for Indian private and PSU bank formats reduces the format-handling time that consumes a disproportionate share of manual review effort.
The seven-step framework above is the minimum structured approach for defensible credit underwriting. Common questions about FOIR thresholds, NACH codes, and authenticity checks are addressed in the FAQs below.