Prove the Household Can Repay — Before You Lend Without Security
Microfinance loans are collateral-free by regulation, so repayment capacity is the only thing underwriting them. Sastra helps India's MFIs evidence household income and surface the obligations a bureau pull never shows.
NBFC-MFIs in India must cap a borrowing household's monthly repayment obligations at a regulated share of its monthly household income, and must assess that income under a board-approved policy. Because the cap applies to the household while credit bureau records are held per individual, lenders increasingly evidence obligations from bank statements. Sastra Technologies builds that analysis for Indian lenders.
What makes microfinance credit different?
No security, and a test applied to the household
Microfinance underwriting isn't a scaled-down version of retail credit. Three features of the segment shape everything:
01
There is nothing to fall back on
Microfinance loans are collateral-free by regulation. There's no asset to value and no recourse to a deposit. The capacity assessment isn't one input among several — it is the whole underwriting decision.
02
The test is at household level, but your data isn't
The repayment-capacity cap applies to the household's total monthly obligations. Credit bureau records are held per individual. Aggregating obligations across a household means either pulling and reconciling records for every adult member — with their identifiers and consent, at a cost per pull — or reading the household's own account for what it actually pays out each month.
03
Ticket sizes make cost per assessment decisive
A per-application data cost that's a rounding error on a ₹10 lakh loan is a real margin line on a ₹30,000 one, and you incur it on the applications you decline as well as the ones you write.
Then there's what the bureau structurally can't see:
- Informal borrowing — moneylenders, shopkeeper credit, family obligations. Never reported, but the repayments leave the account monthly.
- Reporting lag — a loan written last week by another lender may not have reached the bureau yet, and in a segment with this much multiple-lending that gap is where over-indebtedness accumulates.
- Patchy coverage — chit funds, SHG borrowings and some BC-originated loans report inconsistently.
- Obligations of other household members — outside the individual pull you ran.
What the regulation requires of you
The rules that shape microfinance lending in India
The microfinance framework has been revisited since 2022 and the sector has been through a stress cycle — assume nothing is where you last left it. Thresholds are not printed here so the page stays accurate; insert current figures once confirmed.
Household income ceiling
A microfinance loan is defined by being collateral-free and made to a household below a specified annual income threshold. Classification accuracy matters beyond credit: it determines whether the exposure counts toward your qualifying-asset requirement.
The repayment-obligations cap
A borrowing household's monthly loan repayment obligations are capped at a regulated proportion of monthly household income. Meeting it requires you to know the household's total obligations — which is the practical problem this page exists to solve.
A board-approved income-assessment policy
You must assess household income under a policy your board has approved, and be able to show the assessment you made. This is an evidence obligation, not just a process one.
Reporting to credit information companies
Household income and obligation data must be reported, which makes the quality of your own assessment a matter of what you're putting into the shared record.
Qualifying assets
An NBFC-MFI must hold a minimum proportion of its assets as microfinance loans, so borrower classification carries a balance sheet consequence as well as a conduct one.
Pricing transparency
Pricing sits under a board-approved policy with disclosure obligations, including a fact sheet setting out the effective rate and all charges.
How Sastra helps
Find the obligations nobody declared
AB Analyse — bank statement analysis for lenders
AB Analyse reads a household's bank statement and identifies what leaves the account on a schedule: EMIs, recurring transfers to lenders, and repayment patterns that look like debt service whether or not any bureau records them.
For an NBFC-MFI, that does four things:
01
Builds the obligation side of the capacity test — from what the household actually pays, not from what it reported and not only from what was pulled per individual.
02
Exposes multiple-lending early — recurring debits to several lenders are visible immediately, including loans too recent to have been reported.
03
Corroborates declared income — inflows, including benefit and wage credits, checked against what was stated at application.
04
Produces the artifact for your file — a consistent, dated assessment you can put behind your board-approved policy when an auditor or inspection team asks how income and obligations were established.
It works with your field process, not instead of it
Nothing here replaces the field officer or the group. The value is sequencing: the statement corroborates what was declared before the visit, so the officer's time goes to the households where judgement is actually needed, and the interview tests what the data has already flagged. The policy, the decision and the relationship stay yours.
No aggregator subscription, and no per-bank parsers
Cost per assessment. You don't need an Account Aggregator subscription. AA is a clean route to data but it is priced per consent-based pull, and at microfinance ticket sizes that lands hardest — you pay on every application, including the majority you may decline. Reading the statement directly costs materially less per assessment, and works for households not on an AA journey at all.
Coverage. AB Analyse covers 110+ Indian banks because extraction is AI-based rather than a hand-written parser per bank, so an unfamiliar statement layout doesn't need a release. Statements must be unlocked; password-protected files aren't supported.
AB Compliance — RBI returns and audit findings
NBFC-MFIs carry a heavy reporting load: the returns your layer requires, qualifying-asset monitoring, Fair Practice Code obligations and audit findings that need owners and closure dates. AB Compliance automates return preparation and validation, and tracks findings through to resolution.
Customer testimonial
Add a customer quote here in the Drupal editor — a short line on how statement analysis changed household capacity checks before field visits.
The practical questions
Frequently asked questions
Through credit bureau records and, increasingly, bank statement analysis. Bureau data is held per individual and reported with a lag, while the regulatory capacity test applies to the whole household. Recurring debits in the household's account reveal obligations across members, informal lenders and recently written loans.
No. The bureau establishes formal repayment history and is a reporting obligation in its own right. Statement analysis adds what the bureau cannot hold: informal obligations, unreported recent borrowing, and payments made by household members you didn't pull individually.
No. It runs before the visit, so the field officer arrives knowing what the data already contradicts or confirms. Your board-approved policy, your credit decision and your group process are unchanged — the analysis is evidence feeding them, not a substitute for them.
Obligation detection works well even on sparse statements, because recurring outflows are structurally easy to identify. Income reconstruction has less to work with on a largely cash-based account, so treat inflow analysis as corroboration of what was declared rather than a standalone income figure.
No. AB Analyse reads statements directly, which costs materially less per assessment than per-consent AA pulls — you don't pay a data fee on applications you decline. It also covers households that aren't on an AA journey.
AB Analyse is available in two flavours: with the GUI and as an API. The GUI version is available the moment you sign up. For the API version, you will have to write the API calls in your lending application. The time to integrate depends on the vendor.
No. AB Analyse holds no statement data once analysis completes, and the underlying model service runs under an enforced zero-data-retention configuration — nothing is written to durable storage and nothing is shared with the model provider.