Business
Know the Business — Yes Bank Ltd
Figures converted from INR at historical FX rates — see data/company.json.fx_rates. Ratios, margins, and multiples are unitless and unchanged.
Yes Bank is a rebuilt, sub-scale spread bank: the business model is identical to HDFC and ICICI — take deposits, lend the money, pocket the gap — but it operates with structurally cheaper-deposit competitors above it and a still-unfinished CASA franchise. The right way to value the stock is not P/E or fee growth, it is price-to-book × where ROE settles, and the entire bull/bear debate collapses into one question: does ROE migrate from today's 7% toward the peer band of 13-16% over the next three years, or does it stall in the 8-10% zone because the funding-cost gap is structural? The market is pricing roughly the second outcome at 1.36× book — but underestimating two near-mechanical tailwinds (RIDF run-off and rate-cycle deposit re-pricing) that should add 30-50 bps of NIM by FY28 without any operating heroics.
How This Business Actually Works
Yes Bank is a $50.1B balance sheet that converts a 2.6% net interest margin and a growing non-interest income stream into a 0.8% return on assets. Everything else is detail. The economic engine is one stack: pay roughly 5.5-6% for $34.0B of deposits, earn roughly 9.2% on $29.1B of advances, layer fees on top, absorb opex of ~$1.30B (66.7% of income), provision 0.2% of advances for bad loans, and what's left is shareholders'. Each 100 bps of NIM expansion at the current asset base is worth roughly $479M of pre-provision income — more than the entire FY26 net profit.
The spread engine has grown ~50% in four years — but the real shift is the second bar: non-interest income is up ~98% over the same window and is now ~41% of total revenue, among the highest mixes in the peer set. That matters because Yes Bank cannot win the cost-of-funds war against HDFC and Kotak; what it can do is sweat each customer harder for fees, and Q4 FY26 disclosed 41% YoY retail disbursement growth and 21% growth in branch-banking fees, suggesting the cross-sell flywheel is finally working.
Bargaining power sits squarely on the depositor side. Yes Bank pays ~6% on its blended savings balance (cut 150 bps over the rate-easing cycle), and Tonse's team explicitly described liability-side work as "core to a Bank." The reason: in a system with one-tap account portability and HDFC offering equally good UX, retention costs are real. Borrowers, by contrast, are price-takers — repo-linked loans reprice within days of the RBI MPC; deposit rates lag by quarters. This is why a falling-rate cycle compresses Yes Bank's margin in the short run unless CASA accretion and term-deposit re-pricing run faster than asset re-pricing, which is exactly the FY26 management story.
The Playing Field
Yes Bank is the sixth-largest Indian private bank by assets but the cheapest by P/B in the rebuilder cohort, and it sits on the wrong side of the franchise gap. The peer set splits cleanly into three tiers — the premium duo (HDFC, ICICI) earning 14-16% ROE and trading 2.0-2.5× book; the solid #3 (Axis at 13% ROE, 1.83× book); and the rebuilders (Yes, IDFC First, IndusInd) all at sub-8% ROE and trading 1.1-1.4× book. Kotak is the outlier — a small balance sheet but a 4.96% NIM that prices it like a premium franchise (2.1× book) despite only 11% ROE.
The chart is the thesis in one image. Drawing a line through HDFC, ICICI and Axis gives a rough P/B ≈ 0.10 × ROE + 0.7; on that line a 7% ROE company prints at ~1.4× book, vs Yes Bank's 1.36×. The market is pricing the current ROE fairly — the investment question is whether ROE migrates up the curve. Each 100 bps of sustainable ROE corresponds to roughly a 10% multiple step on the peer line, on top of book accretion at the higher rate.
What the peer set reveals about Yes Bank's position. It is closest to IDFC First by every operating metric (similar turnaround vintage, similar mid-cap scale, similar funding-cost handicap), but Yes Bank carries two advantages IDFC First does not: (a) a strategic 20% shareholder in SMBC opening cross-border funding lines, and (b) cleaner asset quality (GNPA 1.3% vs IDFC's ~1.9%, and best-in-class NNPA at 0.2%). It is structurally below Kotak — Kotak's 43% CASA and 4.96% NIM are not catchable in three years; that is a 25-year deposit franchise built around a HNI customer base Yes Bank does not own. The realistic ceiling is Axis-like economics — 13% ROE, 3.5-4% NIM, 40% CASA — not Kotak-like.
Is This Business Cyclical?
Not in the textbook sense — Indian banking demand grew through every recent cycle — but spectacularly cyclical in credit costs, and Yes Bank is the case-study exhibit. The chart below is the entire history. Up to FY18 the bank looked like a high-quality compounder posting 18-21% ROE on aggressive corporate-loan growth. In FY19 the corporate book started to crack (NIM −7%, ROE 6%); FY20 was the credit equivalent of a heart attack — a single year of $4.29B provisions vaporised the equity base, ROE went to negative 68%, and the RBI imposed a Reconstruction Scheme that wiped out AT1 bondholders and brought in SBI as anchor investor. The recovery has been six years of grind back to single-digit ROE.
The cycle hits banks in a sequence: credit costs first, NIM second, loan growth third, capital fourth. Yes Bank lived through all four in 18 months (Mar-2019 to Sep-2020). The current setup is the mirror image — system gross NPAs are at multi-decade lows, Yes Bank's slippage ratio fell to 1.8% in FY26 from 2.1%, retail slippages improved to 3.5% from 4.0% (and Q4 exit was 2.8%), and credit costs are at 0.2%. This is bottom-of-cycle credit and the FY26 P&L benefits accordingly. When the next cycle turns — and history says 4-6 years of benign credit ends, not "ends if X happens" — the most exposed line in the bank is unsecured retail (credit cards + personal loans), where Yes Bank is still rebuilding. Q4 FY26 numbers: $14.2M credit-card slippages, $17.1M personal-loan slippages — manageable today, but the line to watch.
Cycle-aware reading of FY26 numbers. Three of the four positives reported this year ($511M of total recoveries and upgrades — including ~$165M SR write-backs vs ~$128M guidance, 0.2% credit cost, and $30.7M provision write-back from a single resolved corporate account) are bottom-of-cycle artefacts that will not recur indefinitely. Management partially acknowledged this by adding $36.4M of "prudent" contingent provisioning in Q4 — building a buffer for the inevitable normalisation. Strip out the recovery cushion and the underlying core ROA is closer to 0.6%, not 0.8%. The bull case requires core ROA expansion through NIM and opex leverage, not the recovery tailwind.
The Metrics That Actually Matter
Five numbers explain Yes Bank's value creation; everything else is noise. Do not waste time on EPS — the share count has more than doubled since FY20 (capital injections, conversions) and EPS comparisons across years are meaningless. Track these instead:
Why these five and not P/E or ROE? P/E is meaningless for a bank rebuilding off a near-zero base — Yes Bank's 19.9× P/E will halve mechanically as profit normalises, with no change in fundamentals. ROE is the output, not the lever; the inputs are NIM, opex, and credit cost. Of those, NIM and opex are managed-controllable; credit cost is cycle-dependent. CASA and RIDF are the two NIM levers visible before they hit the P&L, which is why analysts who follow Indian banks live on those two slides of the investor deck. The slippage ratio is the early-warning canary that will break before any other line item if the credit cycle turns.
What Is This Business Worth?
Yes Bank is a single-engine bank, not a sum-of-parts: there is one balance sheet, no listed subsidiary worth separating, no holding-company discount, no embedded optionality in a regulated subsidiary. YES Securities (broking) and the wealth platform are tactical fee channels, not standalone valuation lines. The right lens is price-to-book × normalised ROE, with a sense for how fast book compounds at the new ROE. The compact value-driver table below is how a portfolio manager actually frames the under-write — every input is observable, every output rolls up to one number.
Reading the table top-to-bottom. Drivers 1, 2, and 4 are positive and largely mechanical — RIDF runs off whether or not management is good, opex absorption follows from any income growth above 8%, and a 13-15% loan growth target is conservative for an under-leveraged bank with 13.8% CET1. Driver 3 is the negative and inevitable — credit cost normalisation will eat ~80-100 bps of ROE somewhere in FY27-FY29 and the question is only whether the operating leverage (drivers 1 + 2) more than offsets it. Driver 5 is the optionality — SMBC's 20% is the single most under-discussed reason book might re-rate independent of fundamentals; Indian regulatory ceilings make this a multi-year process but the strategic narrative is in place. Driver 6 is the tail you cannot underwrite but cannot ignore.
The number that sets the price. A bank trading at 1.36× book on 7% ROE means the market has already capitalised today's earnings — there is no margin of safety in current numbers. The position is a bet on slope: if ROE migrates to 11% over three years and the implied P/B multiple re-rates from 1.36× to ~1.7×, you earn the multiple expansion (+25%) plus three years of book accretion at 11% retention (+33%) for roughly 60-70% upside. If ROE stalls at 8-9%, the multiple stays put and you earn the book accretion alone — call it 25-30% over three years. If credit costs normalise faster than NIM expands, you can lose 15-20%. Asymmetry favours the long, but the bull case requires patience and the bear case is real.
What I'd Tell a Young Analyst
Do not value this stock on FY26 P/E. A 19.9× multiple on a normalising-from-zero earnings base is meaningless. Build a 3-year forward view of ROE and apply the peer regression P/B ≈ 0.10 × ROE + 0.7. If you cannot articulate why ROE settles above 9%, you do not own it.
Track two slides of the investor deck and ignore the other 60. The CASA ratio slide and the RIDF / mandated-deposits slide. Every 100 bps of CASA or 1 pp of RIDF/assets reduction is worth roughly 7-10 bps of NIM, which compounds straight to book. If both move the right way for four consecutive quarters, the thesis is on track. If either stalls, re-underwrite immediately.
The market is most likely underestimating the cost-to-income trajectory and overestimating how fast NIM closes the gap to peers. C/I fell 460 bps in one year — that is structural operating leverage, not noise, and another 600-800 bps of room exists. NIM, by contrast, is structurally capped 50-100 bps below Kotak/HDFC because the deposit-cost gap is real and CASA accretion is slow. Bull cases that pencil 3.5% NIM in two years are aggressive; 3.0% is realistic.
Watch what management does with the SMBC relationship, not what they say. Cross-border deposit lines, JV announcements, board changes, and any signal of SMBC moving past 20% are the leading indicators. Yes Bank operating in isolation is a 11% ROE bank in 4 years; Yes Bank as the Indian arm of an international franchise is a different multiple.
The thesis breaks on three signals, in order: (1) retail slippage ratio rises through 4% before NIM crosses 3%, (2) cost-to-income flat-lines above 65% for two quarters despite revenue growth, (3) the Supreme Court rules adversely on the AT1 case. If any of these fire, the bull case collapses; if all three are quiet through FY27, the position works.