Strong capital must become durable ROE
Group PAT +26%; bank PAT +29%; group ROE 11.70%. High CET1 protects growth but does not itself earn a premium.
Watch: WATCH: require incremental return improvement
Home / Research / Kotak Mahindra Bank Ltd.
Strong capital must become durable ROE
Group PAT +26%; bank PAT +29%; group ROE 11.70%. High CET1 protects growth but does not itself earn a premium.
Watch: WATCH: require incremental return improvement
Q2 group PAT ₹5,645cr vs ₹4,468cr; bank and lending-related entities ₹4,649cr. Intercompany adjustment −28cr.
Watch: Use consistent group and bank scopes
NII ₹8,459cr vs ₹7,311cr; operating profit ₹6,382cr vs ₹5,268cr; PAT ₹4,212cr.
Watch: Track normalized income and cost
Reported fees 2,926, of which ~158 FCNR processing; related opex ~116. Arithmetic operating profit ex identified net 42 = 6,340.
Watch: Do not remove fees without matching expense
SAGE PRO | Institutional research | 16 November 2026 | Q2 FY27 | WATCH — constructive franchise, conditional on incremental ROE
Evidence: original October 24 Q2 investor presentation and call, accepted and scored in Sage PRO; original exchange statutory filing reference. Amounts in ₹ crore unless indicated. Independent full statutory and date-aligned market/peer valuation reconciliation remains incomplete. Illustrative valuation is not an executable price target.
Key takeaways
Q2 net interest income grew 16% and operating profit 21%, while specific-provision credit cost improved to 35bp from 79bp a year earlier. The balance sheet has a large cushion: bank CET1 21.4% and group CET1 22.1%. But bank ROE was only 11.65% and group ROE 11.70%, materially below what a high-growth franchise might ultimately earn on a more fully deployed capital base.
The variant thesis is that Kotak can turn abundant capital, longer-duration funding and an integrated group into a higher-return franchise through disciplined loan growth, more productive customer acquisition and cost absorption. The special foreign-currency deposit operation provides a fixed-rate funding resource that can support profitable asset deployment.
The counter-thesis is that the bank has expanded its denominator faster than its recurring income, the high capital ratio reflects structural underdeployment, and a larger loan book may bring lower spreads and rising credit costs. The test is incremental risk-adjusted return, not a headline 25% net advances growth figure.
A premium valuation for a conservative balance sheet is defensible only when capital has a productive use or is distributable without damaging the franchise. Excess regulatory capital is not automatically excess cash: risk-weighted assets, operating and market risk, potential expected-credit-loss transition, subsidiary requirements and liquidity constrain redeployment.
We compare Kotak with HDFC and ICICI as larger private-bank franchise references and Axis as an improving returns franchise. The question is not whether Kotak can match each peer's loan growth or NIM but whether its normalized return on equity and growth justify its own valuation.
WATCH until realized earnings on the enlarged balance sheet improve and current price is reconciled.
Key takeaways
AUM is not the bank's own investable asset base. The group model can generate revenue from the customer across deposits, lending, investments, brokerage and insurance, but consolidated financial results include eliminations and different business economics.
| Q2 FY27 group contribution | PAT ₹ crore | Q2 FY26 |
|---|---|---|
| Kotak Mahindra Bank | 4,212 | 3,253 |
| Other lending entities | 437 | 360 |
| Capital market | 428 | 405 |
| Life insurance | 198 | 49 |
| Asset management | 385 | 410 |
| Bank and subsidiaries (including other) | 5,662 | 4,478 |
| Consolidated PAT | 5,645 | 4,468 |
The difference between the sum of bank/subsidiaries and group PAT reflects associates and intercompany adjustments. Group earnings are not the sum of unconstrained cash distributions. The deck states subsidiaries are 100% beneficially owned; still verify regulated capital and actual cash flows before valuing them independently.
The life insurer's current quarter PAT is much higher than a weak prior comparison and should be judged with new-business economics, persistency, investment performance and capital requirements. Capital markets Q2 profit was 428, up 6% YoY but down 31% sequentially.
Asset management profit fell 6% YoY despite group AUM growth of 8%. AUM mix, fees, market levels and operating costs matter; headline AUM growth is not a guaranteed PAT growth rate. Lending entities include Prime, Investments, infrastructure debt and microcredit, which have different borrower and funding profiles.
The group can benefit from cross-sell and customer retention, but intra-group loan transfers need special care. Moving assets from KMIL to the bank increases bank-level advances without necessarily increasing the economic group's loan exposure.
The associated funding and spread may change, but there is no automatic new group customer or new demand. Compare bank and consolidated earnings and capital allocation side by side. In a SOTP model, core bank value should exclude subsidiary value separately added back; do not capitalize consolidated ROE and then add full subsidiary market values on top.
| Standalone ₹ crore | Q2 FY26 | Q1 FY27 | Q2 FY27 |
|---|---|---|---|
| NII | 7,311 | 7,928 | 8,459 |
| Fees and services | 2,415 | 2,500 | 2,926 |
| Other income (total) | 2,589 | 3,338 | 3,194 |
| Net total income | 9,900 | 11,266 | 11,653 |
| Operating expenditure | 4,632 | 5,135 | 5,271 |
| Operating profit | 5,268 | 6,131 | 6,382 |
| Provisions and contingencies | 947 | 668 | 803 |
| PBT | 4,321 | 5,463 | 5,579 |
| Tax | 1,068 | 1,340 | 1,367 |
| PAT | 3,253 | 4,123 | 4,212 |
Key takeaways
Q2 reported fees include approximately 158 of processing fees on FCNR-backed loans; other opex includes approximately 116 of expenses relating to international-branch borrowings. Removing both gives an arithmetic adjusted operating profit 6,340 (6,382−158+116), not an independently established recurring P&L. The matched items indicate why subtracting all special fees without considering their costs would misstate recurring performance. Fee income less the identified processing fees would be 2,768, roughly 14.6% above prior 2,415, assuming the prior base is unadjusted. The full income bridge requires other linked yield and funding effects.
Specific-provision credit cost of 35bp versus 46bp Q1 and 79bp last year supports earnings. However total provisions rose 20% QoQ to 803 from 668. Provisions on advances/receivables 789 included 274 of standard-asset provision versus 77 Q1, of which about 120 related to FCNR-backed advances.
Thus low specific credit cost is not equivalent to all-in provisioning. H1 bank PAT 8,335 vs 6,535 (+28%) and consolidated PAT 11,126 vs 8,940 (+24%). Q2 bank ROA 1.92% vs 1.88% a year earlier; group ROA 2.10% vs 1.97%.
The combination of improved profit and modestly higher ROA is constructive, but quarterly returns on equity remain modest relative to capital capacity.
Illustrative normalized risk-cost sensitivity: on 5,77,295 period-end net advances, a 20bp annual change in risk cost is approximately 1,155 pre-tax per year. A 10bp change in NIM on an illustrative ₹7.5 lakh crore average earning-asset base is 750 annual NII before tax; this is an assumed denominator, not the bank's disclosed earning assets.
Credit loss and margin sensitivities interact through origination mix, and neither can be added to quarterly reported PAT without a consistent time and tax bridge. The base model uses recurring interest and fees, operating cost, specific and standard provisioning and treasury separately.
Key takeaways
The transfer and leverage exclusions measure different effects; they cannot be used to conclude all growth outside the adjusted balance is organic.
Deposits reached 651,491 versus 528,776 a year ago (+23%). The deck identifies 55,344 of FCNR under the swap facility, equivalent to roughly 9.7% of the June deposit base per call. Excluding the amount gives approximately 596,147 and 12.7% YoY growth, consistent with management's 13% underlying deposit growth.
Total average deposits grew 19%, average current accounts 17%, fixed-rate savings 14%, floating-rate savings 4%, and term deposits 22%. End-period CASA was 249,105, with reported CASA ratio 38.2% versus 42.3% last September and 40.3% June.
Ex-FCNR CASA is 41.8% (management disclosure). The distinction matters: the underlying ratio has improved sequentially while reported ratio falls. It is wrong to label the entire CASA-ratio decline as lost customer franchise.
The bank raised roughly US$5.78 billion via the FCNR special window. Management said the resource was used partly to fund the KMIL portfolio acquisition, replace relatively high-cost deposits and reduce reliance on IBPC borrowings, with remaining funds to be deployed over the next few quarters.
International branch borrowings raised during Q2 were approximately 15,800. These are related financing effects, not independent additive liquidity benefits. Analyze balance-sheet funding maturities, cost, swap economics and linked loan exposures before claiming an excess cash pool.
Reported credit/deposit ratio was 88.6% vs 89.4% June and 87.5% a year ago; it should not be mechanically adjusted by subtracting unrelated loan and deposit exclusions without mapping the financing links.
Reported NIM was 4.31% versus 4.54% last year and 4.53% Q1 (22bp sequential contraction). The earnings call identifies a 15bp FCNR impact; adding this back is an arithmetic 4.46% analytical margin, not a separately audited underlying NIM.
The call notes core margin broadly stable after adjustments, but management explicitly does not provide margin guidance. The bank's cost of funds 4.52% vs 4.70% prior year and cost of savings 2.86% vs 2.89% provide a more nuanced view than the decline in headline NIM alone.
The timing of asset deployment and rate repricing determines whether the enlarged funding pool generates an adequate return. Management indicated roughly 60% of the loan book repo-linked in Q&A; fixed and other benchmark exposures should be modelled separately.
A new rate cycle does not improve all assets at once; marginal deposit repricing and duration can offset yield changes.
Key takeaways
No independent product-level cohort data are asserted here. The microcredit subsidiary's Q2 PAT 36 versus loss 22 and H1 63 versus loss 38 demonstrate a recovery from a low base, not proof that microfinance risk has disappeared. Separate standalone bank unsecured exposures from subsidiary results and ensure like-for-like provisioning.
CET1 of 21.4% at the bank (20.9% prior year) and group CET1 22.1% (21.8% prior) are unusually strong relative to the bank's current return on equity. Bank capital adequacy 21.9%, group 22.4%. A large buffer is valuable in stressed markets and can support higher RWA growth, but it depresses financial leverage if not employed.
On a simplified illustrative equity base of 100, a 12% ROE generates 12 annual earnings. Increasing risk-adjusted deployment while preserving asset quality can increase earnings to 15 and ROE to 15%; it is not achieved by merely shrinking the capital denominator.
Actual capital adequacy depends on regulatory risk weights and deductions rather than book equity alone.
Standalone net worth was 145,984, group net worth 194,706 at September. Management highlighted AFS reserve gains not routed through profit, which enlarge equity and depress reported ROE; adjusted bank ROE excluding this non-cash reserve was stated as 12.55% versus reported 11.65%.
This adjusted ratio is useful for interpreting operational return but is not a regulatory capital ratio. AFS volatility and eventual realization may affect equity, capital and liquidity differently. Do not treat the full accounting reserve as freely available funding.
The Q&A also discussed expected-credit-loss transition: management estimated a one-time effect below 2% of net worth, while a previously discussed 12–15bp recurring cost estimate remained subject to clarity. These are management estimates as of October 24, not independently confirmed regulatory or accounting outcomes.
Monitor actual implementation, scope and capital treatment; do not use the estimate as a guaranteed ceiling.
Technology investment and customer franchise growth should be measured by economic output. The call cited cost/assets 2.40% in Q2, declining over six quarters, and underlying sequential expense broadly flat after FCNR adjustments.
The presentation cost/income ratio 45.2% vs 46.8% last year is promising. A bank with growing digital interactions can still face high acquisition and compliance costs. Track net new active customers, stable balances, fee contribution, cost per customer and incident/credit-loss measures.
Productivity is demonstrated by recurring profit per unit of deployed capital, not the number of applications or customer registrations.
Key takeaways
The bank's current returns do not justify an arbitrary high multiple simply because CET1 is high. The investment thesis assumes future return on retained capital can improve.
| Illustrative normalized case | Bear | Base | Bull |
|---|---|---|---|
| Sustainable group ROE | 11% | 14% | 17% |
| Cost of equity | 14% | 12.5% | 12% |
| Long-run growth | 6% | 7% | 8% |
| Specific credit-cost assumption | 0.75% | 0.50% | 0.35% |
| Justified P/B | 0.63× | 1.27× | 2.25× |
The formula (ROE−g)/(Ke−g) is a stylized steady-state clean-surplus valuation with growth below cost of equity. Assumptions are not management guidance or market multiples. The bear assumes funding stays underdeployed and credit losses normalize more sharply, limiting ROE.
Base assumes productive deployment, durable fee income and a measured improvement in operating leverage, with risk costs at a more normal level. Bull assumes higher returns on both retained equity and new assets, profitable subsidiary growth and lower loss rates without aggressive capital release.
At base, one percentage point higher sustainable ROE raises justified multiple approximately 0.18×; a 1pp higher cost of equity lowers base P/B from 1.27× to 1.08×. The terminal growth assumption requires a consistent reinvestment policy; the model is not a forecast of near-term trading returns.
Group book value and ROE are used together for this sensitivity. A separate SOTP may be more appropriate if core bank and group capital deployment differ materially. Under SOTP, value the bank's recurring standalone earnings after removing double-counted subsidiary income, value retained group interests with actual operating/market inputs, and reconcile group net worth, subsidiary capital and eliminations.
Neither current market quote nor peer multiples have been independently reconciled here. An actionable position requires market price, verified equity scope and a conservative assessment of sustainable return rather than a target fabricated from illustrative P/B.
Key takeaways
| Test | Q2 reference | Analyst review condition |
|---|---|---|
| Core net advances | +19% YoY | Sustained growth with risk-adjusted funding |
| Ex-FCNR deposits | +13% YoY | Gap >8pp for two quarters without plan |
| Reported NIM | 4.31% | Below 4.10% for two quarters without mix bridge |
| Specific credit cost | 0.35% | Above 0.70% or deteriorating cohort behavior |
| Bank CET1 | 21.4% | Declining rapidly without return improvement |
| Group ROE | 11.70% | No sustained improvement as resources deploy |
| Fee quality | ₹2,926 crore | Special fees dominate underlying growth |
Thresholds are analyst-defined review signals, not management commitments or regulatory minimums. A positive catalyst is visible loan deployment and stable NIM without weaker underwriting, leading to higher recurring ROE.
A downside catalyst is continuing liquidity drag and rising credit losses as the unsecured/SME mix grows, or a larger-than-expected capital transition effect. The thesis breaks if capital remains underproductive while the bank takes more balance-sheet risk, even if reported growth stays strong.
The existing Sage fundamental assessment (70/100, transcript-only) is retained separately; this report does not create or overwrite a dual-source numerical score. The sector appendix remains unchanged, not counted as company-specific research.
Source register: , document 59403, slides 2–12 (group earnings, net worth, P&L, balance sheet, deposits), and capital/asset quality disclosures. , document 57611, opening remarks, CFO presentation and Q&A (FCNR, KMIL, AFS reserve, ECL and repricing).
Original integrated exchange filing, statutory reference. Accepted source pair verified and scored; independent full financial reconciliation, current-market peer model and detailed ECL impact remain open.
No live price or target asserted.
---
Key takeaways
This is dated research, not a live market quote. Scenarios and historical performance are not guarantees.