Recovery must survive the next credit cycle
PAT +34.2% versus book +11.7%; sequential NIM and asset quality weakened.
Watch: WATCH; validate credit cure and funding repricing.
Home / Research / HDB Financial Services Limited
Recovery must survive the next credit cycle
https://nsearchives.nseindia.com/corporate/HDBF_14102026162735_Intimation_of_Investor_Presentation_sd.pdf
https://nsearchives.nseindia.com/corporate/HDBF_14102026162735_Intimation_of_Investor_Presentation_sd.pdf
https://nsearchives.nseindia.com/corporate/HDBF_14102026162735_Intimation_of_Investor_Presentation_sd.pdf
https://nsearchives.nseindia.com/corporate/HDBF_14102026162735_Intimation_of_Investor_Presentation_sd.pdf
PAT +34.2% versus book +11.7%; sequential NIM and asset quality weakened.
Watch: WATCH; validate credit cure and funding repricing.
11.7% YoY, 2.2% QoQ. Prior balances approximated from rounded growth.
Watch: Compare funded assets, not disbursements alone.
Q2 ₹18,522 crore; growth faster than closing book.
Watch: Track repayments, churn and vintage performance.
NII ₹2,534 crore +15.6%.
Watch: Track comparable average balances and net margin.
SAGE PRO — Institutional equity research | Q2 FY2027 | 16 November 2026 | NSE: HDBFS
Stance: constructive WATCH; valuation-sensitive. Reference close ₹566.40, 13 November 2026 (Sage PRO canonical price history). Research horizon 12–24 months. ₹ crore unless specified. Original issuer/exchange evidence; valuation scenarios are analyst illustrations, not management guidance. Full independent statutory/peer certification is not claimed.
Key takeaways
Q2 FY27 provides evidence of a recovery. Gross loan book reached ₹124,469 crore, up 11.7% YoY and 2.2% QoQ. Disbursements ₹18,522 crore grew 18.7% YoY; net interest income ₹2,534 crore grew 15.6%; PAT ₹780 crore increased 34.2%.
Annualized ROA 2.42% and ROE 14.4%. [A,B] Profit growth faster than assets is encouraging but requires attribution: PPOP growth is lower than PAT growth and credit expense increased sequentially. A strong comparison with last year can coexist with a sequential credit warning.
Our variant view is that the secured enterprise and vehicle franchise, plus a growing consumer distribution engine, can deliver a sustainable recovery if cost discipline and credit selection survive funding repricing. HDB is not the same business as a prime mortgage lender, a gold specialist or Bajaj Finance; the different combinations of secured collateral and unsecured cash-flow underwriting matter.
The portfolio's 73.9% secured mix reduces some loss severity but does not protect all cash flows. We want evidence that new consumer cohorts and geographically diverse vehicle books are contributing to earnings after normal losses.
The counter-thesis is that profitability is recovering at a time when marginal funding is becoming more expensive and seasonal asset-finance stress has increased. The existing low-cost liability book may reprice with a lag, while fixed-rate borrowers cannot necessarily absorb higher rates. If faster consumer originations coincide with weaker collections, earnings may be less resilient than the headline growth suggests.
At ₹566.40, the issuer's Q2 EPS ₹9.40 implies annualized earnings ₹37.60 and ~15.1x run-rate P/E; September book per share ₹266.70 implies ~2.12x dated P/B. These are not TTM or independently audited forward valuation multiples.
The decision depends on whether through-cycle ROE can remain above required return with sustainable book growth. We favor a watch position, not a price-insensitive buy. Require further credit and funding evidence rather than extrapolating one quarter's 34% PAT growth.
| Issuer metric | Q2 FY26 | Q1 FY27 | Q2 FY27 |
|---|---|---|---|
| Gross loan book ₹cr | ~111,432* | ~121,790* | 124,469 |
| Disbursements ₹cr | ~15,604* | — | 18,522 |
| Net interest income ₹cr | 2,192 | — | 2,534 |
| Net total income ₹cr | 2,851 | — | 3,188 |
| Total PPOP ₹cr | 1,530 | — | 1,774 |
| PAT ₹cr | ~581* | — | 780 |
| NIM | 7.95% | 8.35% | 8.23% |
| Gross Stage 3/GNPA | 2.81% | 2.34% | 2.43% |
| Lending cost/income | 40.7% | 39.9% | 39.3% |
| Credit cost | — | 2.32% | 2.37% |
Key takeaways
The two scopes require reconciliation.
NII expanded ₹342 crore YoY, net total income ₹337 crore and total PPOP ₹244 crore. Incremental PPOP represents ~72% of incremental NTI, a useful operating leverage indication. This is not a sustainable incremental margin estimate: expense timing, product mix and the comparison base may matter.
PAT growth of 34% exceeds the 15.9% total PPOP growth, implying changes below operating profit have contributed. A complete statutory bridge requires credit impairment, tax and service-business line items.
Do not attribute the entire recovery to lower credit costs without that bridge.
The sequential picture is different. NIM decreased by 12bp to 8.23%, gross Stage 3 rose 9bp and credit cost rose 5bp. Management described Q1 seasonality and expected asset-finance stress to normalize. The current quarter is better than last year but not uniformly better than June. Investors should resist a narrative that every metric has turned simultaneously.
Credit expense on an annualized basis can move faster than realized write-offs because expected credit loss recognizes risk before recovery. Loan-book growth also affects the denominator of risk ratios.
To evaluate profit sustainability, track the actual credit expense, provisions, recoveries, write-offs and average loan book together. A useful distinction: a lower NPA ratio with higher absolute impaired balances can still mean underwriting risk is rising.
Conversely, higher delinquency during monsoon need not imply permanently higher lifetime loss. The subsequent cure and recovery evidence decides between the two.
ROA of 2.42% and ROE 14.4% indicate positive asset economics with leverage. The annualized EPS/book ratio of ~14.1% uses closing book rather than average equity and therefore differs from issuer ROE. Do not confuse these calculations.
The question is whether the margin of asset-level returns over funding, expense and expected losses is enough to support growing equity without excessive risk. We would not use EBITDA or enterprise value/EBITDA for the lender; net interest income and loss-adjusted equity returns are more meaningful.
Key takeaways
The risk is treating them as homogeneous. Cash-flow lending to a small enterprise depends on business turnover, borrower leverage and collections. Mortgage lending relies on title, property valuation, borrower income and enforceability. Gold relies on custody, valuation and short-duration repayment. A high secured ratio cannot tell us which business earns the best risk-adjusted contribution.
The important investment test is whether the newer cohorts scale without requiring unusually low pricing or weaker standards. Enterprise lending growth is encouraging after a period of caution, but disbursement growth and loan growth are different.
Longer duration, loan churn and repayment affect the closing book. Track the actual funded assets and incremental profit per customer after acquisition and recovery costs. Watch gold-product risk controls as the book grows from a lower base; collateral value can move rapidly while contractual loan balances do not.
A temporary change in gold servicing rules can produce stage migration even if ultimate loss is low.
Commercial vehicles, construction equipment and tractors have physical collateral but depend on operating cash flows. Management reported CV book growth of 8% and construction equipment growth of 12%; seasonal stress was noted in asset finance and expected to normalize.
[B] The economics of a vehicle loan include acquisition/valuation, the financed vehicle's resale price, utilization, freight or construction income, borrower repayment discipline, repossession and auction expense. Collateral is valuable when recoverable, not merely when legally pledged.
A broad distribution platform can help assess local borrower conditions and service collections, but geographic spread does not eliminate a common macro shock. Rising fuel costs can impair operators before aggregate vehicle demand declines.
Weak rural incomes or rainfall effects can delay payments even when collateral remains adequate. A key December-quarter test is not whether vehicle disbursements recover in absolute rupees but whether early-bucket collections, cures and realized credit costs improve on the same cohorts.
Require evidence that seasonal stress has passed rather than automatically classifying the Q2 rise in Stage 3 as harmless.
Consumer finance book grew 22%, disbursements more than 25%, consumer durables book 49%, auto 22%, relationship personal loans 16%. [B] Short-duration consumer-durable finance may provide frequent transactions and customer acquisition; relationship PL can use known customer repayment history, but carries different unsecured risk.
Rapid origination growth can flatter ratios if new loans have not seasoned. Compare loss rates at consistent months on book, approval rates and realized collections, not only portfolio averages.
The company reported a 24.8 million customer franchise, up 18.4%, with gross loan book growth 11.7%. This can mean a broader base and a potentially more granular relationship funnel, but cumulative customer count is not active profitable customers.
Track active borrowers, repeat transactions, share of existing-customer originations, acquisition and service cost, and net lifetime contribution. The issuer indicates the top 20 borrowers comprise ~0.33% of loans.
This is useful individual-name diversification; sector and macro correlations remain.
Management described technology efforts under Shikhar, including AI-assisted collections and onboarding; more than one in five early-bucket telecalls are bot-led. [B p. 3] A bot's call count is an adoption measure, not a profit measure.
Verify cure rate, incremental cash collected, cost per effective contact, customer experience and false escalation. The most valuable technology investment may be an improvement in underwriting or collections rather than a reduction in staff alone.
We do not include a specific AI savings estimate without measured outcomes.
Key takeaways
This implies ~₹175 crore more absolute Stage 3, subject to rounded figures and exact denominator. That is a directional estimate, not a substitute for reported delinquency stocks.
Secured loans were 73.9% of the gross loan book, implying approximately ₹91,982 crore secured and ₹32,487 crore unsecured on rounded weights. Security type, LTV and recoverability differ. A vehicle may be worth less when used markets weaken; a mortgage can require costly legal recovery; a gold loan needs custody and auction controls; unsecured borrower loss depends on cash-flow recovery.
Thus the appropriate risk model segments probability of default, loss given default, recovery timing and collection cost, rather than using a single collateral assumption.
We estimate a rough annual pre-tax sensitivity of ₹311 crore for a 25bp loss-rate move on closing gross loans (₹124,469 × 0.0025). This is not exact issuer average-asset credit cost, but about 10% of annualized quarterly PAT before tax effect.
A shift from 2.37% to 2.8% would be substantial. A stronger economy could lower losses; slower collections or new-vintage deterioration could offset the operating leverage. Stress through-cycle loss assumptions before setting any investment valuation.
The most important alternative hypotheses are: seasonal vehicle-credit deterioration that will cure, versus a lasting increase in repayment pressure or less favorable loan vintages. The call gives management's seasonal interpretation, not verified future outcome.
Monitor gross Stage 2 and Stage 3 additions, customer repayment, restructured balances if applicable, write-offs, cash recoveries and provisioning coverage by vertical. In consumer finance, inspect comparable 3-, 6- and 9-month vintage loss and roll rates; high growth can dilute the overall ratio temporarily.
A thesis breaker is worsening delinquency across multiple new cohorts or persistent asset-finance stress even after seasonal conditions normalize.
Key takeaways
[B pp. 4–6] The difference between average and marginal funding cost is a real risk: refinancing at higher costs can change average liabilities gradually, even if current NIM is stable. Conversely, higher loan yields or product mix may offset costs; reset frequency, competitive pricing and prepayment determine effectiveness.
Reported cumulative ALM mismatch was positive across buckets up to five years and capital adequacy was 21.37%. [B] These are favorable issuer-reported indicators; they are not an independent liquidity stress or maturity reconciliation.
HDB's parentage and funding access should not be interpreted as access to bank deposit economics at no incremental cost. Lenders must account for refinancing concentration, asset duration and liquidity needs.
Reported capital adequacy is a buffer, not unrestricted deployable equity.
A 25bp annual margin change applied to closing gross loans is about ₹311 crore pre-tax on a crude proxy. Together with a 25bp credit-cost shock, this implies ~₹622 crore annual pre-tax sensitivity. The model uses closing loans, not the exact average assets or loan-book denominator.
It illustrates that relatively modest joint changes can materially affect earnings. Evaluate capital and credit costs alongside margin: higher-yield products are not necessarily higher-return when loss and servicing cost rise.
Peer economics should be compared through a matched framework. Bajaj Finance has a larger multi-product consumer and cross-sell platform, while Chola and Shriram provide stronger vehicle-finance comparisons.
Tata Capital offers a diversified lender comparison with a different housing and motor finance mix. A raw P/E or GNPA ranking is insufficient because consolidated scopes, credit definitions, leverage, return periods and capital transactions differ.
HDB's reported 14.4% ROE is below the headline reported returns of some larger diversified peers but this is not a same-basis valuation proof. Compare through-cycle ROA, capital adequacy, normalized credit losses, funding cost, attribution to shareholders and sustainable book growth.
Do not infer that a lower quoted multiple means cheapness without the normalized earnings comparison.
Key takeaways
A simple terminal residual-income model P/B = (ROE − growth)/(cost of equity − growth) under steady-state clean surplus gives ~0.86x for ROE 12%, cost 13%, growth 6%; 1.6x for ROE 15%, cost 12%, growth 7%; 2.5x for ROE 18%, cost 12%, growth 8%. These are illustrative terminal economics, not actual current fair values; an explicit high-growth period can justify a higher current multiple.
Our next-year normalized EPS model starts from 4 × ₹9.4 = ₹37.6, not an audited FY27 forecast. Bear assumes −12% from the annualized run-rate and 11x multiple, reflecting margin and credit pressure; base +12% and 15x, requiring stable risk-adjusted spread and growth; bull +25% and 18x, requiring stronger operating and credit conversion. Multiples are analyst assumptions, not observed peer P/E.
| Scenario | Growth vs annualized Q2 | EPS ₹ | Assumed P/E | Value ₹ | Return vs ₹566.40 |
|---|---|---|---|---|---|
| Bear | -12% | 33.09 | 11x | 363.97 | -35.7% |
| Base | 12% | 42.11 | 15x | 631.68 | 11.5% |
| Bull | 25% | 47.00 | 18x | 846.00 | 49.4% |
Formula: annualized Q2 EPS × (1 + growth) × multiple. Price return excludes dividend, transaction costs and future capital changes. These are scenario sensitivities, not probability-weighted price targets. The model deliberately acknowledges that earnings may decline from a strong current quarter. A more rigorous institutional model would forecast average asset book, yield, funding cost, expense and credit costs separately and reconcile to attributable diluted shares.
Book sensitivity: if book per share grows 12% annually for two years while P/B contracts from ~2.12x to 1.7x, price return before dividends is only ~0.6%; at constant multiple it is 25.4%. The implication is that sustainable book growth can be offset by a revaluation of risk. Entry valuation discipline is therefore inseparable from asset quality and capital. No independently matched peer P/E model is asserted.
| Test | Q2 evidence | Analyst monitoring trigger |
|---|---|---|
| NIM | 8.23%; 8.35% Q1 | Stay around 8.2%; investigate sustained compression below 8% |
| Credit cost | 2.37%, up from 2.32% | No broad-based sustained deterioration; reconcile loss expense |
| Gross Stage 3 | 2.43%, vs 2.34% June | Absolute stock and ratio stabilize; verify actual cures |
| Asset finance | Seasonal stress disclosed | Collections/cures improve after monsoon; state and vintage evidence |
| Consumer | Book +22%, durables +49% | Comparable vintage loss and contribution remain healthy |
| Funding | Average ~7.04%, NCDs ~8.2–8.25% | Assess rollover and floating asset repricing, not current rate alone |
| Capital | CAR 21.37%; ROE 14.4% | Stable prudential buffer and improving asset-level return |
Key takeaways
Interim dividend ₹2/share was declared with 23 October record date; do not annualize automatically.
The thesis breaks if NIM falls below the expected range while credit expense rises, if fast consumer/enterprise cohorts show persistent weaker underwriting, or if asset-finance stress fails to cure across geographies. It strengthens if margins hold despite liability repricing, absolute Stage 3 stabilizes with cash recoveries, and growth comes from profitable seasoned products. No allocation recommendation is implied; use a suitable risk budget and reassess normalized valuation.
https://nsearchives.nseindia.com/corporate/HDBF_14102026162735_Intimation_of_Investor_Presentation_sd.pdf
https://nsearchives.nseindia.com/corporate/HDBF_20102026124950_Intimation_of_Transcript.pdf
https://nsearchives.nseindia.com/corporate/HDBF_20102026124950_Intimation_of_Transcript.pdf
https://nsearchives.nseindia.com/corporate/HDBF_20102026124950_Intimation_of_Transcript.pdf
This is dated research, not a live market quote. Scenarios and historical performance are not guarantees.