EASE 9.0 · Socio-Economic Impact · MSME & Priority Sector Credit

Collateral Was Never the Real Signal

Meeting priority-sector targets without relaxing portfolio discipline.
Underwriting on cash flow, not collateral alone.

Priority-sector lending targets exist for a reason, but meeting them by relaxing portfolio discipline is a false choice. An MSME with genuine repayment capacity, evidenced by consistent GST filings and healthy cash flow, is a good credit risk regardless of collateral.

Collateral-Based Underwriting Misses the Real Signal

A small manufacturing unit with three years of steady GST turnover and disciplined current-account cash flow can be turned down for a working-capital loan simply because it lacks the immovable-property collateral a traditional model requires, while a business with weaker cash flow but strong collateral sails through.

The Path to Progress

The problem was never that these businesses are riskier. It is that manual underwriting cannot economically assess cash-flow-based creditworthiness at scale.

The PSU Precedent, and the Market Most Banks Ignore

Punjab National Bank offers a live precedent for the PSU peer set, not a hypothetical one. It became the first public-sector bank to go live on GST Sahay, an OCEN- and Account-Aggregator-based app built by SIDBI with the RBI Regulatory Sandbox, that finances individual sales and purchase invoices in real time against GST data, with no manual document collection. That is a working, in-production example of exactly the cash-flow-first approach this page argues for, not a private-fintech capability PSU banks are still catching up to.

The sharper, less obvious point is what happens after a loan is correctly tagged. RBI's own e-Kuber platform runs a live Priority Sector Lending Certificate (PSLC) market: a bank that over-originates in one priority-sector category can sell the surplus, and a bank that falls short can buy compliance rather than distort its lending book to chase a quota. The market settled ₹498 billion in its first year. That reframes the strategic question this category poses: priority-sector lending is not only a retail-underwriting problem to solve with better scoring, it is also a treasury decision about whether to originate where cash-flow-based AI gives the bank a genuine edge, and buy PSLCs where it does not, rather than treating the target as a single lending-volume number to hit by any means.

Sources: SIDBI/GST Sahay coverage (ThePrint, 2023); Business Standard, “PNB launches app to enable MSMEs access instant loans using GST invoices” (2023); Arthapedia/GKToday summaries of RBI's Priority Sector Lending Certificate framework.

The Business Case

The tension in priority-sector lending is not between meeting regulatory targets and maintaining quality; it is between the cost of manual assessment and the volume needed to meet targets profitably. That trade-off carries a real, RBI-defined penalty attached to getting it wrong: a bank that falls short of its priority-sector targets must park the shortfall in NABARD's Rural Infrastructure Development Fund at a quasi-penal rate, up to four percentage points below the Bank Rate for the largest shortfalls, a standing rule that triggers automatically, not a hypothetical drag on returns. A bank that automates cash-flow-based underwriting can meet its priority-sector obligations from a segment of MSMEs it previously could not economically assess at all, rather than diluting standards on the segment it already served, absorbing that RIDF penalty for falling short, or paying to buy PSLCs it could have originated more cheaply itself.

Source: RBI Master Circular on Priority Sector Lending, RIDF contribution provisions.

Where AI Agents Fit

Role: Enabler

Cash-flow scoring is backend work an MSME borrower never sees directly; they simply get approved or not, on different evidence than before. A Credit Decisioning Agent configured for this segment perceives business signals rather than personal ones: GST filings, current-account cash-flow patterns, and invoice or receivable data where available, the same category of data GST Sahay and the TReDS exchanges already move in production. It scores repayment capacity against a policy calibrated for business cash flow rather than collateral value, and a Compliance Agent verifies the loan is correctly tagged against the specific RBI priority-sector sub-category it satisfies at the moment of sanction, rather than reconstructed from a portfolio review months later. That reconstruction is how most banks discover a shortfall against target, usually after it is too late in the reporting cycle to correct it and too late to decide whether buying a PSLC would have been cheaper than the collection effort.

The Skeptic's Answer

The sharpest pushback here concerns data integrity: GST filings and current-account statements are still self-reported and gameable in a way a registered mortgage on immovable property is not. Collateral requirements exist, the argument goes, because cash flow can be dressed up for a loan application in a way a title deed cannot.

Start with where the numbers actually come from. GST Sahay and the Account Aggregator rails pull filings and transaction history directly from GSTN and the bank's own core systems, cryptographically signed and consent-based, rather than a PDF a borrower could edit before submitting. And collateral valued once at origination does not update; a cash-flow score re-derived from live GST and banking data every reporting cycle catches deterioration a static mortgage valuation never would. The PSLC market is the backstop for the residual risk: originate where the AI gives a genuine edge, buy certificates where it doesn't, rather than pretending cash-flow scoring alone closes every gap collateral used to cover.

One Underwriting Engine, Tuned to a Different Segment

The same policy-bound, auditable decisioning engine that runs a bank's mainstream corporate credit book runs cash-flow-based MSME underwriting, with signals, thresholds, and priority-sector tagging configured for this segment's data profile and regulatory categories. What changes is the question a treasury desk answers every quarter: originate the shortfall, buy a PSLC, or park it in RIDF at four points below the Bank Rate. Only one of those three actually grows the book.