Axis Bank
The improving ROE story
- Last close
- ₹1,212.10
- 29 Sept 2026 · reference
- 1D · 1M
- +0.2% · −4.2%
- price-only
- Weight
- 6.0%
- 31 Jul 2026 · Aug rank 2
- Thesis review
- 8 Sep 2026
- Why We Own, p28
The improving ROE story — what has to happen, what we believe, what breaks it
- 2QFY27 NIM (the trough call) and the quantum of FCNR(B) deposits raised.
- Retail loan growth crossing 10% and cost-to-assets printing ≤ 2.15%.
- Any RBI relaxation on unsecured risk weights or a benign ECL transition disclosure.
- The numbers are still improving: ROE goes from 12.7% (FY26) to 14% (FY27E) to 15.6% (FY31E) on our model as NIM recovers off the June-2026 trough, cost-to-assets falls below 2.15% and credit cost normalises to 65 bp — exactly the pattern the market has paid for historically.
- Cheaper than ICICI and Kotak: 1.7x core book against a 1.7x fair value from our residual-income model, with an 18% base-case and 16% probability-weighted 12-month return (BUY).
- Liability franchise catching up: deposits +18% YoY in 1QFY27, CASA 38–40%, the FCNR(B) scheme is bringing in cheap dollar funding, and the retail term-deposit engine built through 2024–25 is now lowering marginal cost.
- Operating leverage: opex growth held at 9–10% against 13–14% revenue growth; the Citi integration costs are behind it.
- House history: we owned Axis through its credit-cost trough over the past two years (up 40–50% for us) and the July file shows it at 6.0% — the second-largest position.
- Core · Large cap6.0% of PMS · rank 2
- NIM: Axis has a history of margin disappointment; if 2Q NIM falls below 3.40% (reported) the recovery thesis is delayed by a year and the stock would test 1.5x book.
- Unsecured retail: credit cards and personal loans are ~10% of the book; a consumer-credit cycle would lift credit cost above the 65 bp we model.
- Deposit competition: LDR of 92% is above ICICI's 89%; a return of the FY25 funding squeeze would cap growth at 12% rather than 16%.
Thesis and position rationale
- Investment case
- The improving ROE story
- Why this business
India's third-largest private bank (₹12.6 lakh crore of loans, ~5,900 branches) with a corporate-bank heritage that has been rebuilt into a balanced franchise since 2019: retail and SME are now ~70% of loans, the Citi consumer business (2023) added a premium card and wealth franchise, and the bank has the strongest technology delivery among the big three after ICICI. Subsidiaries (Axis Finance, Axis AMC, Axis Capital, Axis Securities, and a 20%-plus stake in Axis Max Life) are worth ~₹110 a share.
- What we believe
- 01The numbers are still improving: ROE goes from 12.7% (FY26) to 14% (FY27E) to 15.6% (FY31E) on our model as NIM recovers off the June-2026 trough, cost-to-assets falls below 2.15% and credit cost normalises to 65 bp — exactly the pattern the market has paid for historically.
- 02Cheaper than ICICI and Kotak: 1.7x core book against a 1.7x fair value from our residual-income model, with an 18% base-case and 16% probability-weighted 12-month return (BUY).
- 03Liability franchise catching up: deposits +18% YoY in 1QFY27, CASA 38–40%, the FCNR(B) scheme is bringing in cheap dollar funding, and the retail term-deposit engine built through 2024–25 is now lowering marginal cost.
- 04Operating leverage: opex growth held at 9–10% against 13–14% revenue growth; the Citi integration costs are behind it.
- 05House history: we owned Axis through its credit-cost trough over the past two years (up 40–50% for us) and the July file shows it at 6.0% — the second-largest position.
- Why now
- See sector chapter and catalysts.
- Market disagreement
- ICICI Bank: We own both; ICICI is the higher-quality, higher-priced finished article. Axis has more upside if the ROE recovery lands, and more risk if NIM disappoints again.
- Kotak Mahindra: Kotak's 12% ROE at 2.0x core book compares with Axis's 14% at 1.7x — Axis is cheaper for a higher and rising return.
- IndusInd Bank: IndusInd is a governance and accounting rebuild after the derivatives loss; Axis has already done its rebuild.
- Position sizing
Core Large cap 6.0% of the PMS on $31 Jul 2026 (August rank 2). Core positions are owned through the cycle for leadership and cash-flow quality.
- Catalysts
- 2QFY27 NIM (the trough call) and the quantum of FCNR(B) deposits raised.
- Retail loan growth crossing 10% and cost-to-assets printing ≤ 2.15%.
- Any RBI relaxation on unsecured risk weights or a benign ECL transition disclosure.
- Risks and response
- NIM: Axis has a history of margin disappointment; if 2Q NIM falls below 3.40% (reported) the recovery thesis is delayed by a year and the stock would test 1.5x book.
- Unsecured retail: credit cards and personal loans are ~10% of the book; a consumer-credit cycle would lift credit cost above the 65 bp we model.
- Deposit competition: LDR of 92% is above ICICI's 89%; a return of the FY25 funding squeeze would cap growth at 12% rather than 16%.
- Thesis-break conditions
- Not stated separately on this page; the risk list carries the monitoring triggers.
- Review history
- 8 Sep 2026 · Yash Palod · one-pager in "Why We Own What We Own" (p28) · portfolio as of $31 Jul 2026
- 31 Aug 2026 · Classification in the August top-30: Core
Source: Buoyant_Why_We_Own_What_We_Own_Sep2026.pdf (portfolio as of 31 July 2026), p28. Internal; external publication of these fields is controlled by audience policy.
Internal actions
- HoldPosition carried into August at rank 2.
- ReviewNext scheduled: post 2QFY27 results (Oct–Nov 2026).
Add/trim/exit decisions require transactions data; none supplied. Recording a decision needs a persistence adapter (not configured).
Sector datapoints
- System deposit growth 15.4% (Aug-26); credit +18.3%; corporate bond issuance −18% is pushing wholesale borrowing back to banks
- FCNR(B) scheme has brought $20 bn+ into the system, easing dollar funding for the large private banks
- Bank Nifty below long-run P/B average; Axis at 1.7x core is at fair value — only HDFC Bank (1.6x) is cheaper among the big four