> ## Content Index > Fetch the complete content index at: https://research.finbox.in/llms.txt > Use this file to discover other available public pages before exploring further. # The Pattern #226: The RBI now wants to see who you lent to - URL: https://research.finbox.in/newsletter/the-pattern/the-pattern-226-the-rbi-now-wants-to-see-who-you-lent-to/ - Published: 2026-11-06T09:58:01.000Z - Updated: 2026-11-06T09:58:01.000Z - Description: The new directions for priority sector lending comes with a new transparency guideline for banks doing PSL through NBFCs. - Author: Mayank Jain - Tags: #ThePattern, mayank jain, RBI, public sector lending, RBI guidelines, Loan Origination Software, NBFCs, Digital Lending Hello everyone, Welcome to the 226th edition of The Pattern, a weekly newsletter on the latest in finance, technology, and the economy. On Thursday, 5 November, the RBI released [five draft directions](https://econotimes.com/India-s-RBI-Seeks-Public-Input-on-Five-Priority-Lending-Drafts-1771621?ref=research.finbox.in) that consolidate its priority sector lending rules, one for each type of bank. Most of what's in them is housekeeping. Education loans stay eligible up to ₹25 lakh, agri-infrastructure gets a [₹100 crore per-borrower limit](https://www.theweek.in/wire-updates/business/2026/11/05/rbi-proposes-rs-100-cr-limit-on-bank-loans-for-agri-infra-under-priority-sector-lending.html?ref=research.finbox.in), and comments are due by 26 November. One section goes further. When a bank lends to an NBFC so that the NBFC can lend on to farmers or small businesses, the [draft for commercial banks](https://rbi.org.in/Scripts/bs%5Fviewcontent.aspx?Id=5205&ref=research.finbox.in) says the bank can count that loan towards its priority sector target only once the NBFC has passed the money on to the end borrower. To prove it, the bank must collect six things about every one of those borrowers: name, amount sanctioned, amount outstanding, tenure, disbursal date and PSL category. The NBFC has to flag each loan in its own systems so that auditors and RBI supervisors can check it. That's what I want to talk about in this edition: what happens to a wholesale loan once the lender can see through it. **A rule that used to stop at the balance sheet** On-lending is the quieter, older cousin of co-lending. The bank lends the NBFC a lump sum, the NBFC lends it on in small tickets, and the bank books it as priority sector credit. The [current directions](https://www.rbi.org.in/Scripts/BS%5FViewMasDirections.aspx?id=12799&ref=research.finbox.in) cap this at 5% of a bank's priority sector lending. They ask banks to keep "disaggregated data" on NBFC portfolios and "borrower-wise details" for housing finance companies. As far as I can tell from the current text, they don't say which fields, and they don't ask the NBFC to tag anything for supervisors. The draft spells out both. [Vinod Kothari's team](https://vinodkothari.com/2026/11/rbis-psl-norms-consolidation-important-change-for-psl-linked-lending-to-nbfcs/?ref=research.finbox.in), who read these rules more closely than most, flagged this as the important change for PSL-linked lending to NBFCs. The money around it isn't small. Bank credit to NBFCs stood at [₹20.65 lakh crore in March 2026](https://www.careratings.com/uploads/newsfiles/1778152239%5FFY26%2520Credit%2520Offtake%2520Closes%2520Strong%2520NBFCs%2520at%2520the%2520Forefront.pdf?ref=research.finbox.in), 9.7% of all bank credit, and had grown 26.3% in a year. Only a slice of that is PSL on-lending, but it's a slice banks have a particular reason to want. **Why a bank pays for paperwork** Commercial banks have to direct [at least 40% of adjusted net bank credit](https://www.rbi.org.in/Scripts/BS%5FViewMasDirections.aspx?id=12799&ref=research.finbox.in) to priority sectors. A bank that falls short can buy priority sector lending certificates, or it gets allocated deposits in funds like NABARD's RIDF, which pay between Bank Rate minus 2 and minus 4 percentage points depending on the size of the shortfall. So a loan that counts as PSL is worth more to the bank than its interest rate alone. My read is that some of that value reaches the NBFC as cheaper or easier funding. That part is my inference, and the draft says nothing about pricing. Say an NBFC borrows ₹500 crore from a bank to lend to small enterprises. The per-borrower limit for that kind of on-lending is ₹20 lakh, so that's at least 2,500 borrowers. If the average ticket is closer to ₹5 lakh, it's 10,000. Under the draft, each of those 10,000 loans needs its six fields with the bank and a flag in the NBFC's system, and the money has to stay deployed in PSL assets. "Amount outstanding" changes every month, so the record has to keep up. If the file falls behind, the bank's PSL claim on that ₹500 crore gets hard to defend in front of an auditor. **The see-through loan** The bank is now buying two things from the NBFC: a borrower who repays, and evidence that the money went where the NBFC says it went. **A wholesale loan used to buy the bank a balance sheet. Now it buys the bank a borrower list.** I'd call this the see-through loan. It's credit extended to an intermediary but underwritten, in part, on the intermediary's ability to show its working. Once the bank's benefit depends on proof that lives inside the NBFC's systems, the quality of those systems becomes part of the NBFC's cost of funds. Two NBFCs with the same credit quality could get different terms from the same bank, because one can produce a clean, flagged, current file and the other sends a spreadsheet three weeks late. ![](https://storage.ghost.io/c/88/cf/88cfcfc1-f936-46a1-a0db-77c479da9277/content/images/2026/11/The-Pattern--226---The-see-through-loan_1.png) **The see-through loan: what a bank needs to see before PSL on-lending counts, before and after the 5 Nov draft.* Everyone here is behaving sensibly. The RBI wants to know the money reached a farmer and not a corporate treasury. The bank wants its PSL credit without an argument with its auditors. The NBFC wants cheap money without handing a customer list to a lender that may want those customers too. Add those up and the cheapest bank money starts drifting towards the NBFCs that can prove where it went. Those overlap with the NBFCs best at reaching the borrower, but they aren't always the same firms. Co-lending reached this point from the other side. Since 1 January, [each partner must hold at least 10%](https://www.business-standard.com/finance/news/rbi-revises-co-lending-norms-mandates-10-loan-retention-loss-guarantee-125080601927%5F1.html?ref=research.finbox.in) of every co-lent loan, and the bank's share sits on its own books, so it sees the borrower by default. I looked at what that did to NBFC economics in [#222](https://research.finbox.in/newsletter/the-pattern/can-nbfcs-keep-co-lending-profitable/). The draft gives on-lending some of that visibility, while the bank still holds none of the individual loans. **Why I might be wrong** The strongest case against me is that this is a draft, and mostly a consolidation. Borrower-wise details for HFCs, disaggregated data for NBFCs and auditor certificates against double-claiming are all [in the current rules](https://www.rbi.org.in/Scripts/BS%5FViewMasDirections.aspx?id=12799&ref=research.finbox.in). NBFCs already report every loan to the credit bureaus [every fortnight](https://www.rbi.org.in/scripts/NotificationUser.aspx?Id=12718&ref=research.finbox.in), so most of the six fields already exist somewhere. The PSL tag is the new part. The cap also limits the stakes. Five per cent of a 40% target works out to roughly 2% of a bank's credit, and lending to MFIs that belong to an RBI-recognised SRO sits [outside the cap](https://rbi.org.in/Scripts/bs%5Fviewcontent.aspx?Id=5205&ref=research.finbox.in) altogether. If banks treat the six fields as a once-a-year audit exercise, pricing won't move, and I'll have over-read some paperwork. **Where it gets fixed** The data mostly exists, but in the wrong place. NBFCs send loan-level data to all four credit bureaus every fortnight, while the bank needs the same rows, with a PSL tag, for its own auditor. Nobody owns a shared pipe for that, and each bank's auditor will want evidence from its own records. So every NBFC risks building the same file again for each lender. The second constraint is trust, because the NBFC is handing beneficiary names to a lender that may compete for them. If I ran treasury at a mid-sized NBFC, I'd put PSL tagging on the same list as cost of funds. I'd find out how quickly I could hand a bank a flagged, current file for every on-lent rupee, and I'd bring it up in my next funding conversation before the bank did. If I ran a bank's PSL desk, I'd use the consultation window to push for one standard format, so I'm not reconciling ten versions of the same file. **What I'm taking away** In [#212](https://research.finbox.in/newsletter/the-pattern/the-rbi-just-made-source-of-truth-a-regulatory-requirement/), I wrote about the RBI's draft data governance norms and its push for a single source of truth inside each lender. This draft carries the same instinct across the boundary between two lenders. The proof now has to travel with the money. Until it travels as easily as the money does, the NBFCs with the cleanest records will keep getting better terms than the NBFCs with the best reach. Before I sign off, a question for you. If you handle bank borrowings at an NBFC, could you produce those six fields, flagged and current, for every on-lent loan this week? Just write back. I'd especially like to hear from NBFC treasury teams and from anyone on a bank's PSL desk. Until next week, Mayank If this was useful, forward it to one person who works on bank borrowings at an NBFC, or on priority sector compliance at a bank. *Views are personal and do not reflect those of FinBox or its promoters.* **Reading list** - [The Pattern #222: Can NBFCs keep co-lending profitable?](https://research.finbox.in/newsletter/the-pattern/can-nbfcs-keep-co-lending-profitable/) (FinBox, 9 Oct) - [The Pattern #212: The RBI just made "source of truth" a regulatory requirement](https://research.finbox.in/newsletter/the-pattern/the-rbi-just-made-source-of-truth-a-regulatory-requirement/) (FinBox, 17 Jul) - [State of Digital Lending 2026: a FinBox report](https://research.finbox.in/guides/state-of-digital-lending-2026-a-finbox-report/) (FinBox, May 2026) - [RBI proposes ₹100 crore limit on bank loans for agri infra under priority sector lending](https://www.theweek.in/wire-updates/business/2026/11/05/rbi-proposes-rs-100-cr-limit-on-bank-loans-for-agri-infra-under-priority-sector-lending.html?ref=research.finbox.in) (PTI, 5 Nov) - [RBI's PSL norms consolidation: important change for PSL-linked lending to NBFCs](https://vinodkothari.com/2026/11/rbis-psl-norms-consolidation-important-change-for-psl-linked-lending-to-nbfcs/?ref=research.finbox.in) (Vinod Kothari Consultants) - [Gold loan book triples since 2023, delinquencies rise sharply in 2026](https://business-standard.com/amp/finance/news/india-gold-loan-delinquencies-transunion-cibil-portfolio-august-2026-126110501124%5F1.html?ref=research.finbox.in) (Business Standard, on TransUnion CIBIL's report, 5 Nov) - [Cheap money is running out, and India's banks know it](https://www.thehansindia.com/business/market-compass/cheap-money-is-running-out-and-indias-banks-know-it-1137976?ref=research.finbox.in) (The Hans India, 1 Nov) - [UPI transactions rise 5.8% MoM to record 25.47 bn in October](https://inc42.com/buzz/upi-transactions-rise-5-8-mom-to-record-25-47-bn-in-october/?ref=research.finbox.in) (Inc42, 2 Nov)