The Narain Call — where the house view gets challenged
Aditya Narain mentoring call, Sep-2026 (written up 6-Sep-2026). Former head of research, Citi India; for a long stretch the most-followed sell-side voice on Indian banks.
What it means for the fund, in one sentence
~₹21,000 crore across the PMS and four AIFs; roughly 55% large cap, 20% mid, 15% small, 10% cash, run core-and-satellite where the edge is drawdown control and the ability to exit at scale.
Hold banks as a diversified, liquid beta on an India catch-up trade, size them for a 10–15% move rather than a re-rating to old means, and if the goal is upside leverage to a market rally, look at the platform businesses the banks ceded.
Narain's comments do not argue against owning banks; they argue for being precise about why we own them and what we expect from them. What the framing does NOT support is telling distributors that private banks are cheap relative to their own history and will compound back to it.
On broker consensus: Nobody is wildly bullish or bearish. ICICI is everybody's top pick for CASA, growth and credit cost; HDFC Bank is liked but with merger-related patience; target prices sit 20–25% above spot, which is roughly what the same brokers expect from the Nifty. The sector call and the market call are the same call.
1. FCNR(B) changes the sector's RISK, not its return
Each bank is hedged in rupees, so no individual bank carries currency risk. But the country now has a defined dollar liability of roughly $120–130bn sitting inside a headline reserve number nobody nets off. In a good cycle the repayment is forgotten; in a bad credit or currency cycle it becomes the story and bank stocks are punished for sitting in the middle of it even though they are hedged. His phrase: it raises the beta on banks.
Contradicts the briefing: The briefing treats the FCNR window as the funding bottleneck breaking — an unambiguous positive.
2. 19% loan growth is not being rewarded, and the market is right
Part catch-up after five thin years, part inventory and working-capital financing that is not investment. When commodity prices rise a dealer needs twice the working capital for the same tonne and bank credit rises with no real activity behind it. A steadier 16–17% for five or six years would be healthier than 19% for two. Corporate credit will probably be better from here, but not in a way that earns banks a higher multiple.
Contradicts the briefing: The briefing reads 19% system credit growth as a structural volume story at 1–1.5x nominal GDP.
3. Mean reversion is the weakest part of the bull case
“The biggest fallacy is the mean reversal you will hear plenty of as far as private banks are concerned. Those means also change. HDFC Bank traded at five, six times book for years, and it will never get back there, even if the economics look the same.” ROE has drifted from 17–18% to 14–15% and the economy is a 12–14% credit-growth economy, not a 20% one. Nothing about that is cyclical, so nothing about it reverts. Expect a 10–15% catch-up driven by top-down India sentiment — an oil price fall, for instance — and not much beyond that.
Contradicts the briefing: The briefing anchors fair values on banks sitting in the lower half of their own historical P/B band.
4. CASA is drifting, not collapsing
Cash-management technology means nobody with real money leaves it in a savings account, so the old 43–44% peaks are gone. But the technology has existed for fifteen years and the ratio is still around 40% system-wide, so a fall to 10–15% is not coming either. The challenger playbook of buying deposits with 6–8% savings rates is finished. Nothing structural to model; watch individual franchises instead.
Contradicts the briefing: The briefing treats CASA as cyclical with a structural tail and does not name what is pulling money out.
5. Part of the derating is structural — platforms took what banks were meant to own
Banks were supposed to own India's platform economy: they had the customers, the accounts and the payment rails. They lost it to brokerages, asset managers and wealth platforms. Those businesses are P&L-driven and leveraged to a market rally; banks are balance-sheet-driven and are not. If the market runs 10%, the brokerages run 30% and the banks run 12–14%. That is where he would look for beta, not in the banks. ICICI Securities was founded 35 years ago; a decade-old discount broker made it irrelevant.
Contradicts the briefing: The briefing attributes the derating to foreign selling and cyclical margin compression.
Where the call disagrees with the briefing
The house briefing was written 31-Jul-2026; the call was September. Where they conflict, this is the more recent view.
| Topic | Buoyant briefing (31 Jul 2026) | Narain (Sep 2026) |
|---|---|---|
| Funding | Funding is the binding constraint; CD ratio near 80% limits growth | System is now over-funded for 2–4 years; the constraint is the RBI absorbing liquidity, and roll-off risk in a bad cycle |
| Loan growth | Structural volume story at 1–1.5x nominal GDP; corporate cycle 'if real' | 19% is partly catch-up and partly working capital; 16–17% sustained is the healthy case; corporate better but no boom |
| Valuation anchor | Every covered bank in the lower half of its historical P/B band; fair values 12–40% above spot | Historical bands are the wrong anchor; means have reset with ROE and growth; 10–15% top-down upside, then nothing bank-specific |
| Cause of derating | Flow-driven (foreign selling) and cyclical margin compression | Partly structural: platforms captured the disintermediation banks were meant to own; sector structure static for a decade |
| CASA | Cyclical with a structural tail; competitor for deposits unnamed | Gradual drift from cash-management convenience; no cliff; challenger rate-war playbook is over |
| Rates | The debate is cut versus hold | The RBI skipped the rate response it used in 2013; absorption is via liquidity tools; softer-than-otherwise rates near term |
| Top risk | Funding; unsecured re-acceleration; corporate concentration | Higher sector beta from the FCNR(B) liability if the credit or currency cycle turns |
What to watch from here
A monitoring list different from the briefing's — these are the things that tell you which scenario is playing out.
- RBI liquidity absorption
Size and frequency of variable-rate reverse repos, and where call and overnight rates settle relative to the repo. Absorption that keeps pace is the benign path; rates drifting well below the policy rate says the RBI is behind.
- Net reserves, not gross
Headline reserves less the FCNR(B) liability. If the RBI spends reserves defending the rupee while the liability is unchanged, the market will eventually make the same subtraction.
- The rupee's response to the inflow
It barely moved on $120–130bn. Continued weakness despite the money would confirm the market is already netting the liability, and would raise the sector's beta further.
- The shape of loan growth, not the level
Whether the 18–19% print settles toward 16–17% or fades faster; the split between investment, working capital and refinancing; the share coming from gold-backed and other secured retail.
- Whether the market starts paying for growth
If bank multiples begin to respond to growth prints, the catch-up trade is under way and should be sized as a 10–15% move. If they do not, the structural argument is winning.
- Oil, as the trigger for the India trade
A fall toward $70 is the kind of external event that starts the top-down move banks would benefit from.
- Early asset-quality signals two to three quarters out
A negative turn is the trigger for banks to stop deploying the FCNR(B) money and park it with the RBI — the moment the repayment starts to matter for valuations.
- CASA at the system level against the 40% mark
A gradual drift is expected and is not news; a sharp break would be, and would most likely coincide with a strong equity market pulling savings into platforms.
- Platforms versus banks
If brokerages, asset managers and wealth platforms outrun banks by the two-to-three-times ratio he described in a rising market, the structural-derating argument is being confirmed in real time.