The landscape is a tower. You only watch the lit floors.
Since October 2022, RBI sorts every NBFC into a four-layer tower — Scale-Based Regulation. Size and systemic importance decide the floor; the floor decides how hard RBI watches them. Your first act of judgement is knowing which floors even matter to a bond investor.
The Scale-Based Tower
Tap a floor to read what lives there.
Upper Layer — 14 systemically important giants, verified name-by-name against the 31-Mar-2026 RBI registry, re-listed by RBI every year, facing bank-like scrutiny. The safest end of your NCD universe lives here.
Your investable universe is the two lit floors. Upper Layer for the anchor positions, Middle Layer for the yield — and the deeper you go into the Middle Layer, the harder every later post on this page must work.
The Watchlist · the names to know
My earlier caption on this page said "roughly 15" — a memory-based estimate. The actual 31-Mar-2026 RBI registry, checked row by row, shows exactly 14. Thirteen of the names above were already right; one had simply been renamed — Aditya Birla Finance is now Aditya Birla Capital Limited. Kept visible rather than silently fixed: this is what "verified" should mean on this page from here on.
The complete register runs to thousands of names and refreshes quarterly — any copy pasted here would quietly rot, and a Watchtower must never trust a stale map. This page names the recurring cast; the living lists stay one tap away:
RBI master lists (layer-wise & category-wise): rbi.org.in → BS_NBFCList
Industry mirror of the same lists: fidcindia.org.in → List of NBFCs
RBI's batch cancellations (35 in Dec 2025; 150 + 7 surrenders in May 2026; 135 + 13 in Jun 2026) are overwhelmingly Base Layer shells — defunct paper companies, mostly West Bengal and Delhi. The failures that hurt bondholders — IL&FS, DHFL, SREI — surfaced first in rating actions and ALM disclosures, long before any registration action. Watch the right feed (Post 05), not the loudest one. Also note: a November 2025 amendment exempts entities with no public funds and no customer interface from registration entirely — the tower's ground floor keeps shifting.
Every NBFC is the same machine: borrow, lend, lose, spend — keep the rest.
Before any P&L makes sense, hold the machine itself. An NBFC borrows money at one rate, lends it at a higher rate, loses some to borrowers who don't pay, spends some running the shop — and what survives is profit. Four numbers. That's the entire business. Run it yourself:
The Spread Machine
Per ₹100 of loans, funded ₹85 by debt and ₹15 by equity (the ~15% capital floor). Move the four levers.
A healthy machine — this is the zone where durable NBFCs live.
ROA of 1.5–2.5% is where good NBFCs live; leverage turns it into 12–18% ROE. Anything richer means one of your sliders is set to a fantasy — usually credit cost, and usually it corrects three years later.
Notice what the machine teaches: a lender can manufacture growth instantly — approve more loans, book more interest income today. The credit cost of those loans arrives in year two and three. That single time-lag is why every red flag in Post 03 exists.
Read the strategy the way a lender to the lender would.
You are not buying the company's dream; you are lending it money. So read its growth strategy with one question: does growth make my coupon safer or riskier? Three disclosures answer it — all free, all public.
The three strategy disclosures & what to extract
- The loan book mix (investor presentation, first few slides). Secured vs unsecured share, retail vs wholesale, ticket size, top-20 borrower concentration. Granular, secured, retail = boring = good for a bondholder.
- The borrowing mix (presentation + annual report). Banks, NCDs, commercial paper, deposits, ECBs. Diversified and long-tenor = resilient. Heavy short-term CP funding long loans = the IL&FS shape.
- The ALM statement (annual report notes / rating rationale). Cash inflows vs outflows bucket by bucket. This is your standing lesson from the ledger: IL&FS and DHFL died of ALM mismatch, not just bad loans.
Red flags vs green flags
The free cheat sheet: the rating rationale. Every rated NBFC has a public rationale on CRISIL, ICRA, CARE or India Ratings — a trained analyst has already read the book mix, the ALM statement and the strategy, and written down exactly what worries them. It is the single highest-value free document in this entire manual. Read it before the glossy investor presentation, never after.
Six lines tell you everything. The rest is furniture.
An NBFC's quarterly results look dense; the machine from Post 02 is hiding inside them in plain sight. Here is the waterfall, top to bottom — with what each line whispers to a bondholder:
| Interest income | − | The lending yield in rupees. Rising faster than AUM? The book is shifting riskier. |
| Interest expense | = | Cost of funds. Rising faster than peers'? Lenders are quietly repricing the risk. |
| Net Interest Income | + | The raw spread. The engine's torque. |
| Fee & other income | − | Nice, but never lend against it. |
| Operating expenses | = | Cost-to-income ratio lives here; branch-heavy models run higher. |
| Pre-Provision Profit | − | The shock absorber. This is what stands between bad loans and your coupon. |
| Impairment (ECL) | = | Credit cost. The line management is most tempted to understate. |
| Profit before tax | What survived. Compare to PPOP: a thin gap means provisions are eating the cushion. |
The vital signs beside the P&L
Quarterly results & presentations → BSE/NSE corporate announcements · Full ALM & ECL notes → annual report · The analyst's synthesis → rating rationale · Bond-level terms & security → placement memorandum / trustee report.
Four signals, in order of earliness.
Distress announces itself in a fixed sequence. Watch the feeds in this order and you hear the news years apart — watch only the last one and you hear it after the queue has formed.
Downgrades, outlook changes to Negative, and watch placements. IL&FS went AAA → D in weeks, but the outlook chatter started earlier. One downgrade is information; two in a year is a verdict.
crisilratings.com · icra.in · careratings.com · indiaratings.co.in — search issuer name → latest rationaleListed issuers must disclose interest delays, defaults, resignations of auditors/directors, and results — on the exchange, often before the news covers it.
bseindia.com → Corporates → Announcements · nseindia.com → Companies → Announcements — filter by issuerDebenture trustees publish security-cover certificates and default status for each ISIN — this is where "senior secured" gets audited against reality. NSDL's bond database carries ISIN-level terms.
indiabondinfo.nsdl.com — ISIN lookup · trustee sites (Catalyst, Axis Trustee, IDBI Trusteeship, SBICAP) · sebi.gov.inCoR cancellations and surrenders (monthly-ish batches), penalties, PCA invocations, board supersessions. Confirmation, not prophecy — by the time RBI supersedes a board, feeds 1–3 screamed long ago.
rbi.org.in → Press Releases · rbi.org.in → BS_NBFCList (registered / layer lists, refreshed quarterly)Surveillance is a habit, not a project.
The two-cycle coupon test taught you the pattern: small, repeated verification beats occasional heroics. Same discipline here. Checkboxes reset when the page closes — screenshot the completed list into your monthly records, the same way you keep the coupon ledger.
One row per candidate. No row, no deal.
This is Gate 3 of your five gates, made physical. A candidate that cannot fill this row cleanly does not proceed to price talk. The UGRO row is started from its public listing — finish it yourself with the 3-source rule before it counts. Tap any shaded cell to type; screenshot the finished card.
| Field | UGRO Capital · INE583D07661 | Next candidate |
|---|---|---|
| Layer / type | NBFC · verify layer on RBI list | |
| Coupon · maturity | 9.75% fixed · 26-Nov-2027 | |
| Seniority · security | Senior Secured (per listing) · charge on…? cover…x? pari passu? | |
| Rating · trajectory | Ind-Ra A+ / Positive (per listing) · pull rationale + history | |
| GNPA / NNPA trend | ||
| CRAR | ||
| ALM comment | ||
| Borrowing mix | ||
| Book mix (sec/unsec) | ||
| Verdict · gate 3 |
Price transparency → yield reality → this scorecard → portfolio fit & three-floor framework → the behavioural gate. Issuer intelligence sits in the middle for a reason: a fair price on a failing issuer is still a bad deal.
The tower stands only if you can rebuild it from memory.
Answer aloud before revealing. Return tomorrow, then next week — the retrieval is the construction.
1. Which two floors of the tower contain essentially your entire investable NCD universe — and roughly how many entities live on each?
2. State the Spread Machine in one sentence — the four levers and what's left over.
3. Why is fast AUM growth the easiest number for a lender to fake — and when does the truth arrive?
4. Rank the four feeds by earliness — and which one is confirmation rather than prophecy?
5. Which line of the P&L is the shock absorber between bad loans and your coupon — and which line is management most tempted to understate?
6. What is the single highest-value free document on any rated NBFC — and when do you read it?