FULL PUBLISHED RESEARCHEarnings and management commentary analysis
16 November 2026 | Earnings period: 30 September 2026 | WATCH — conditional constructive bias
The case for SBI rests on a large, relatively low-cost deposit franchise, improving domestic spreads and a better-capitalised balance sheet. The case against an aggressive earnings extrapolation is that two successive September quarters contain major stake-sale profits. The underlying operating comparison is much less impressive than the headline; the next leg of the investment thesis must come from recurring income, sustainable asset quality and a profitable conversion of credit growth.
This research uses the original 3 November analyst presentation and call, with the original exchange filing linked for statutory reference. Financial amounts are in ₹ crore unless otherwise stated. The bank's presentation definitions are used; this is not an independent full statutory reconciliation.
1. Investment thesis: own the franchise, not the exceptional gain
Reported Q2 net profit of 23,038 grew 14.28% and operating profit of 34,226 grew 7.28%. NII rose 14.69% to 49,143. These three growth rates tell different stories: interest income is expanding strongly, while the non-interest/cost bridge limits operating leverage and lower provisions support bottom-line growth. The headline profit includes 7,288 from the SBI Funds Management stake sale. The previous Q2 contained a 4,593 YES Bank stake-sale profit. On the management presentation's operating-profit definition, excluding the respective gains produces 26,938 versus 27,311, a 1.37% decline. This is an analytical adjustment, not an audited alternative earnings measure. It leaves treasury valuation and other unusual expenses in both periods.
Key takeaways
- The differentiated thesis is not that SBI has become structurally less profitable.
- Rather, a low 0.22% Q2 credit cost and income from disposals should not be capitalised as permanent annual earnings.
- Domestic NIM of 3.11%, up 11 bps sequentially, gives a plausible recurring-income route to better earnings.
Gross advances grew 17.84% while deposits grew 12.27%. A durable positive view requires the funding cost of that gap to remain manageable. An investor should separate three clocks: quarterly NIM repricing, the next two to three quarters of expense timing and deposit mobilisation, and the longer-term transition to expected-credit-loss accounting and capital consumption.
SBI is the scale benchmark among PSU banks but should not be valued merely by a PSU discount. The group owns substantial financial-services businesses, including stakes in listed companies. The right test is sustainable return on retained common equity and capital allocation, with separate subsidiary value only where it is not already reflected in the core bank valuation.
Public-sector mandate and funding breadth are both strengths and constraints. A large branch network can attract stable savings balances and cross-sell products; it also imposes a recurring cost base that cannot be assumed away when digital usage rises.
Maintain WATCH until pricing, funding and earnings normalization jointly support a better risk-adjusted entry.
2. Operating performance: growth is real, quality of earnings needs a bridge
The Q2 presentation (slide 19) provides the following whole-bank financial series.
Key takeaways
- NII growth was driven by interest income rising 10.46% while interest expense grew 8.09%.
- Fee income of 9,549 versus 8,711 rose 9.63%; fee growth is more relevant to the recurring thesis than investment-disposal receipts.
- Reported investment sale/revaluation income of 6,011 versus 7,470 includes the effect of the disposal and treasury results.
Management explained in the call that the MF sale flowed through profit while a treasury mark-to-market loss of about 1,275 offset part of the benefit. It also explained that NSE investment disposal is treated through the AFS reserve rather than P&L. Neither should be added indiscriminately to ordinary operating earnings.
The presentation's disclosed gain adjustment subtracts 7,288 and 4,593 from the respective operating profits. A mechanical add-back of the current treasury loss would raise adjusted current profit to about 28,213, but would not establish a comparable growth rate unless prior-period treasury movements are consistently normalised. Therefore the base thesis uses the conservative stake-sale-only bridge, not a cherry-picked treasury add-back.
The operating expense increase of 9.63% also deserves scrutiny. Staff expense rose from 16,606 to 17,452 and overheads from 14,393 to 16,531. Finance management said about 850 of expenses were booked upfront to spread eligible expenditure across quarters instead of a Q4 bulge.
This is timing, not a permanent saving. H1 expenses rose 7.64%, H1 NII 14.78%, and H1 net profit 12.31%. Reported H1 cost-to-income fell to 48.33% from 48.53%; Q2 cost-to-income rose to 49.82% from 49.28%.
A rough Q2 ratio calculated on operating income stripped only of the disclosed stake gains is 55.78% versus 53.16%. This is not the bank's published ratio and includes remaining non-recurring items, but it highlights the hurdle for actual core efficiency.
A successful Q3/Q4 should demonstrate better recurring revenue absorption, not just a lower reported cost ratio when investment profits rise.
3. Liability franchise, margins and credit deployment
Total deposits reached 62,77,919, compared with 55,91,700 a year earlier. CASA balances rose 8.13% to 22,96,883 and retail term deposits rose 16.74% to 27,85,198. The presentation shows domestic CASA ratio of 38.32%. The faster growth of term balances is not inherently bad if asset yields compensate for the higher funding cost; it is a warning against assuming that SBI's historical average liability cost applies to incremental lending. Domestic credit-to-deposit ratio rose from 69.82% in September 2025 to 73.02% in June and 74.23% in September 2026. Whole-bank liquidity coverage was 140.47%, up from 126.05% in June and below 143.05% a year earlier. These are different measures: a comfortable LCR does not mean a lender can permanently grow loans faster than stable deposits without affecting marginal spreads.
Key takeaways
- Quarterly domestic NIM recovered from 3.00% in Q1 to 3.11%, versus 3.08% last Q2.
- Whole-bank NIM is 2.94%, compared with 2.86% Q1 and 2.96% a year ago.
- H1 domestic and whole-bank NIMs were 3.06% and 2.90%, respectively.
In the cumulative yield/cost chart, domestic advance yield fell from 8.68% to 8.22% over the year and domestic deposit cost from 5.13% to 4.84%. This is compatible with the loan repricing and deposit reset cycle, but these cumulative ratios cannot be mixed directly with a single quarter spread. The margin recovery in Q2 is encouraging; two more quarters of comparable numbers are needed before treating it as a new equilibrium.
Credit growth is diversified rather than dependent on one product. Domestic gross advances were 43,77,012, up 16.94%, and foreign offices 8,31,332, up 22.86%. Domestic segments at September were personal retail 18,25,083 (+14.54%), agriculture 4,51,438 (+22.64%), SME 6,61,335 (+21.97%) and corporate 14,39,156 (+16.12%).
RAM represented 67.12% of domestic advances. Management reiterated 14–15% full-year credit growth in the call. The difference between the 17.84% year-on-year September snapshot and that full-year guidance is important: seasonal credit deployment and base effects make a straight-line extrapolation inappropriate.
Track corporate drawdowns, agriculture seasonality and SME risk-adjusted pricing together. A diversified book reduces single-sector dependence, but the faster-growing SME and agriculture pools require stronger early-warning surveillance.
An illustrative sensitivity: each 10 bps annualised change in spread on 50 lakh crore of earning assets changes annual NII by about 5,000 crore, before mix, timing and hedging. This is a scale illustration, not a forecast of SBI's earning-asset denominator or margin beta.
It explains why modest deposit repricing errors can outweigh fee growth and why quarterly loan growth should be assessed with a marginal funding model. Key decision variable: Q3 domestic NIM above 3.05% with deposit growth catching up is more valuable than another headline gain in profits.
4. Asset quality, capital and the expected-loss transition
Gross NPA ratio fell to 1.39% from 1.73% last year and 1.47% in June. Net NPA was 0.36% versus 0.42% last year. Gross NPAs declined in absolute terms from 74,272 in June to 72,529 in September, so improvement is not just denominator growth. Q2 fresh slippages were 5,227 versus 7,046 in Q1 and 4,754 last Q2. Gross additions of 5,377 included 150 increases in outstanding balances. Total reductions were 7,120, of which recoveries and upgrades accounted for 3,235. The remaining reductions should not be assumed to be cash recovery; the presentation distinguishes the categories. The closing bridge is 74,272 + 5,377 − 7,120 = 72,529. The better sequential slippage print coexists with higher fresh slippages year on year; both observations matter.
Key takeaways
- Segment NPA ratios at September 2026 were agriculture 6.56%, SME 2.87%, personal 0.65%, corporate 0.78% and international 0.11%.
- Agriculture is the highest-risk ratio among disclosed pools and has grown 22.64%; the next year should test whether this combination is sustainable after seasonal collection and renewal cycles.
- SME absolute NPAs rose slightly from 18,921 in June to 18,977 even as its ratio fell from 2.93% to 2.87%.
This is a useful denominator-effect warning. Overall PCR was 74.43% versus 75.79% a year ago; PCR including AUCA was 91.83%. AUCA-inclusive coverage is not equivalent to current balance-sheet coverage. The Q2 credit cost of 0.22% is below 0.39% a year ago. For earnings valuation use a normalized credit-cost assumption rather than perpetuating the exceptionally benign quarter.
Capital adequacy rose to 16.09% from 14.62% last year; CET1 was 13.34% versus 11.47%, according to slide 26. Capital support includes changes in eligible plough-back under the revised framework, as well as earnings and reserve movements.
Credit RWA to gross advances fell to 52.11% from 55.99%. Do not assume risk weights will keep falling as the book grows. The call explicitly highlighted uncertainty about expected-credit-loss implementation and parallel run: management did not quantify volatility or the eventual capital effect.
Treat the 1 April 2027 implementation discussion in the call as management commentary, not a separately verified regulatory deadline. Watch transition disclosure, regulatory capital deductions, incremental provision estimates and post-transition CET1 headroom.
A rising CET1 ratio today is not a guarantee against a future model-driven provision or capital adjustment.
5. Productivity and group optionality: evidence before valuation
SBI's scale makes small efficiency improvements meaningful. The presentation reports 98.9% of transactions through alternate channels, 11.1 crore registered YONO customers and roughly six crore registrations on New YONO. In the call, management said more than 1.2 crore Re-KYC updates were completed digitally, branch intervention fell more than 50%, approximately 50,000 man-hours per day were saved and over 9,500 Seva Sarathis deployed. These are reported activity/productivity measures, not audited monetary savings. Redeployed staff hours could support service quality, deposit mobilisation or lending throughput without immediate expense reduction. The test is the relationship between incremental fee/NII revenue, staff cost per unit of business, customer acquisition cost and reported overheads over the next several quarters. Cyber resilience and fraud prevention costs must be included in that assessment. Digital scale is a moat only if it improves risk-adjusted returns after operating costs.
Key takeaways
- The group adds economic optionality.
- Slide 40 reports stakes of 68.58% in SBI Card, 55.31% in SBI Life and 55.56% in SBI Funds Management.
- Their H1 FY27 PATs were 1,442, 1,274 and 1,566 respectively, with reported ROEs of 17.50%, 12.87% and 46.81%.
The deck also identifies SBI General Insurance (73.87% stake, H1 PAT 205) and SBI Capital Markets Group (100%, H1 PAT 816). These figures are those published by the subsidiaries; accounting bases and minority interests differ. They cannot simply be added to solo-bank earnings. A sum-of-parts model should use current separately verified market values for listed stakes, realistic valuation ranges for unlisted businesses, tax/holding-company considerations and a core bank equity value excluding the value being added back.
No contemporaneous market prices are used here, so a numerical SOTP price target would be false precision. The MF partial disposal demonstrates monetisation value, but also reduces future economic participation relative to the pre-sale holding. Evaluate the proceeds and ongoing stake on separate lines.
6. Valuation and scenarios: sustainable ROE is the anchor
Solo capital plus reserves at September 2026 was 599,586 crore (923 capital and 598,663 reserves), versus 518,059 a year earlier. This book-equity arithmetic is not adjusted tangible book and includes accounting reserve effects. H1 reported ROE of 18.68% and ROA of 1.14% should not be equated with forward normalized profitability. A useful stylised bank valuation is justified P/B = (sustainable ROE − long-run growth)/(cost of equity − long-run growth). This assumes a steady-state clean-surplus relationship and growth below the cost of equity, and is not a short-term fair value model. Use a common 5% long-run nominal growth assumption for scenario comparability.
Key takeaways
- All scenario inputs are analyst assumptions, not management guidance.
- Bear: domestic NIM below 3%, term funding expands faster than CASA, credit cost returns above 0.50%, transition provisions constrain capital and recurring operating profit weakens.
- Base: NIM near 3.05–3.15%, mid-teens lending growth with funding convergence, credit cost normalises in a 0.30–0.45% band, and core efficiency gradually improves.
Bull: NIM sustains above 3.15%, profitable credit deployment and improved liability mix preserve capital generation, credit losses remain low without releases and digital productivity improves recurring efficiency. The bull ROE is intentionally close to reported H1 rather than assuming perpetual exceptional gains.
These multiples describe a range of assumptions, not a buy/sell recommendation at the current market quote. Compare the market's actual P/B with forward sustainable return estimates and verify equity scope before applying them.
Use subsidiary SOTP as a cross-check, not as a second layer of value added to a multiple already capitalising group earnings. Compare SBI with Bank of Baroda, Canara Bank, Punjab National Bank and Union Bank on normalized ROA, credit cost, liability mix and capital, rather than ranking solely by reported P/E.
ICICI and HDFC offer private-sector reference points, but funding mix, operating cost structure, subsidiaries and governance differ. No peer multiples are asserted without a date-aligned verified market-data snapshot.
7. Decision, catalysts and monitoring checklist
Investment decision: WATCH — conditional constructive bias. The bank has improving quarterly domestic spreads, substantial liquidity, broadly distributed credit growth and a stronger capital buffer. But stake-sale-adjusted operating profit is slightly lower year on year, Q2 reported cost-to-income rose, term deposits are outgrowing CASA, and fresh slippages are higher than last Q2. The right position-size decision depends on entry valuation and portfolio risk; it cannot be inferred solely from earnings growth.
Key takeaways
- Three positive catalysts: another quarter of domestic NIM at or above 3.10% without a funding squeeze; deposit growth moving closer to advances growth with stable CASA mix; and positive stake-sale-neutral operating-profit growth with expense timing clearly reconciled.
- Three downside catalysts: accelerated term-deposit pricing and domestic NIM below 3.00%; SME or agriculture slippage deterioration despite favourable aggregate ratios; and a material unanticipated ECL/capital transition charge.
- As analyst monitoring thresholds rather than guidance, use domestic CD ratio above 78%, CET1 below 12%, Q3 fresh slippages materially above the Q1 7,046 level, and a sustained credit cost above 0.50% as prompts to revisit the thesis.
Review after the next disclosed results; use actual results dates, not an assumed release calendar. Do not interpret any threshold as a regulatory limit or a price stop.
The base thesis is invalidated if margin and core operating earnings deteriorate together while funding needs and credit costs rise. A better entry case requires the market price to offer a discount to conservative normalized book-return economics, with risk sized to the investor's horizon. Preserve the shared sector report below as context; its cross-bank conclusions do not override SBI's own quarterly numbers.
Source register and verification limits
- Original Q2 FY27 investor presentation, 3 November 2026: BSE attachment, slides 6–9, 13–19, 26–30, 40–45; document ID 69633.
- Original Q2 FY27 analyst call transcript, 3 November 2026: BSE attachment, management opening remarks and Q&A on earnings normalization, treasury, expense timing, credit growth and ECL; document ID 86069.
- Original exchange financial result reference: NSE integrated filing, period ending 30 September 2026. Link retained; independent line-by-line XBRL reconciliation remains incomplete.
- The accepted source pair was scored in Sage PRO. This report separates direct disclosed figures, derived arithmetic and analyst assumptions. Current market price, date-aligned peer valuation, independent statutory reconciliation and exact ECL capital effect are not verified here. No target price or full independent certification is claimed.
---
Banking sector peer research
Key takeaways