Profit recovery is real, driven by operating income and lower impairment, but the microfinance cycle and rapidly expanding retail finance require…
Turnaround with disciplined risk limits
Watch: Monitor catalysts and falsifiers
Home / Research / CreditAccess Grameen
Profit recovery is real, driven by operating income and lower impairment, but the microfinance cycle and rapidly expanding retail finance require separate risk underwriting.
https://www.bseindia.com/xml-data/corpfiling/AttachHis/9d90680b-2fb2-43a4-8818-9ec848ab6353.pdf
https://www.bseindia.com/xml-data/corpfiling/AttachHis/9d90680b-2fb2-43a4-8818-9ec848ab6353.pdf
https://www.bseindia.com/xml-data/corpfiling/AttachHis/9d90680b-2fb2-43a4-8818-9ec848ab6353.pdf
https://www.bseindia.com/xml-data/corpfiling/AttachHis/9d90680b-2fb2-43a4-8818-9ec848ab6353.pdf
Turnaround with disciplined risk limits
Watch: Monitor catalysts and falsifiers
Issuer-reported figures and explicitly identified derived comparisons; see chapter analysis.
Watch: Compare next quarter using identical definitions.
Issuer-reported figures and explicitly identified derived comparisons; see chapter analysis.
Watch: Compare next quarter using identical definitions.
Issuer-reported figures and explicitly identified derived comparisons; see chapter analysis.
Watch: Compare next quarter using identical definitions.
Research date: 16 November 2026 | Price reference: ₹1102.6 (13 November 2026)
Stance: Turnaround with disciplined risk limits. Core thesis: Profit recovery is real, driven by operating income and lower impairment, but the microfinance cycle and rapidly expanding retail finance require separate risk underwriting.
Key takeaways
The group-loan book ₹23,746 crore declined about 2% since the FY27 start due to write-offs and graduation. The key question is whether a hybrid group/individual lending platform can diversify earnings and improve borrower lifetime value without recreating over-indebtedness.
Management raised FY27 ROA guidance to 5.2–5.8% and ROE to 21–24%, with AUM growth guidance 20–25% and credit cost 3–4%. These are expectations, not outcomes. H1 PAT ₹995 crore already exceeds FY26 full-year ₹777.6 crore, showing the scale of cyclical recovery. The 13 November share price ₹1,102.60 must be judged against normalized through-cycle earnings rather than annualizing the first clean quarters indefinitely. [1,2]
| ₹ crore | Q2 FY26 | Q1 FY27 | Q2 FY27 |
|---|---|---|---|
| PPOP | 694.8 | 872.5 | 970.2 |
| Impairment | 525.7 | 212.5 | 299.2 |
| PBT | 169.2 | 660.0 | 670.9 |
| PAT | 125.8 | 493.4 | 501.6 |
| NIM | 13.3% | 14.4% | 15.1% |
| Cost/income | 32.5% | 29.3% | 28.2% |
Key takeaways
This explains why PAT rose just 1.7% QoQ despite strong NIM improvement. Credit cost cannot be ignored in extrapolating the quarter. H1 PPOP ₹1,842.7 crore vs ₹1,347.8 crore; impairment ₹511.7 crore vs ₹1,097.5 crore, producing H1 PAT ₹995 crore. Interest income de-recognized on Stage 3 fell from ₹76.9 crore to ₹37.8 crore YoY in Q2; this improvement supports NIM but is partly cyclical. Recoveries ₹18.7 crore versus ₹16.4 crore are separately reported. [1]
Key takeaways
Group loans historically rely on social and field collection infrastructure; graduating established customers to individual business and secured loans can improve retention and loan sizing. The danger is transferring overextended borrowers to larger loans before their cash flows justify it. Retail AUM per retail borrower is roughly ₹1.06 lakh (₹7,200 crore / 6.8 lakh); group AUM per remaining 38.2 lakh about ₹62,200, assuming mutually exclusive borrower categories. The assumption may not hold exactly due to overlapping loans. This suggests a larger ticket risk as well as opportunity.
Secured retail book ₹695 crore, +60.9% YoY, only about 9.7% of retail finance and 2.2% of total AUM. Hence most of the ₹7,200 crore retail segment is not yet secured. A headline 'retail diversification' must not be confused with a dramatic reduction in unsecured credit risk.
Borrower retention rose 85% FY26 to 88% H1 FY27. Digital collections 26.3% from 18.1% year ago can improve payment traceability but do not remove household cash-flow risk. The graduation strategy is validated only if repeat customers have lower realized losses after controlling for loan size.
[2]
Key takeaways
The ₹137 crore macroprudential buffer and 3.31% ECL provisioning are relevant protections, but a reserve increase also represents real uncertainty. H1 write-offs ₹614 crore, predominantly group loans. Low current NPA must be reconciled with prior write-offs; write-offs reduce measured gross NPA but don't prove cure.
Flooding and rainfall in Bihar, UP, Chhattisgarh and MP caused higher PAR accretion in August/September; drought-related requests for more time emerged in parts of Maharashtra in October. Local stress can spread when rural income and borrower overlap deteriorate.
A 100-bp annual loss increase on ₹30,946 crore AUM is about ₹309 crore pretax, roughly 12% of annualized Q2 PBT ₹2,684 crore. A 200-bp shock is ₹619 crore. Given unsecured borrower economics, this stress is more relevant than simply extrapolating the recent GNPA ratio.
Track new PAR generation, roll rates, cure, write-offs and vintage loss by state and product. [1,2]
Key takeaways
A 50-bp yield decline on ₹30,946 crore AUM is approximately ₹155 crore annual pretax if applied universally, before pricing mix and funding offsets. The company aims to balance competitiveness with loss, expense and funding cost through its pricing policy.
Equity ₹8,919 crore, CRAR 26.9%, Tier 1 26.2%. October signed $175 million ECB, and FY27 domestic NCD issuance ₹940 crore. These diversify funding tenure but create currency-hedging, refinancing and counterparty considerations; do not treat a signed ECB as September cash.
Marginal and average borrowing cost both around 9.1% per management; changes in funding mix can move the marginal rate independently of benchmark shifts. Balance-sheet cash and bank balances ₹1,576 crore, investments ₹1,925.9 crore at September.
Combined liquidity indicators are substantial but exact usable liquidity requires encumbrance and maturity analysis. [1,2]
Key takeaways
These imply indicative equity values of ₹10,533 crore, ₹23,174 crore and ₹33,708 crore; not per-share targets or observed peer multiples. The 13 November market price ₹1,102.60 requires the audited diluted share count to compare precisely.
Book equity ₹8,919 crore is the second anchor: annualized Q2 PAT/equity ~22.5%, roughly consistent with management's raised ROE range. A stress-return of 15% on the same equity gives PAT ₹1,338 crore, illustrating why a microfinance credit cycle can substantially change valuation.
A higher equity multiple needs sustained above-cost-of-equity returns after a normalized 3–4% annual credit cost. The most important downside sensitivity is a simultaneous NIM reduction and regional loss increase, not merely slower AUM.
[1,2]
| Case | Annual PAT ₹ crore | Assumed multiple | Equity value ₹ crore |
|---|---|---|---|
| Bear | 1505 | 7x | 10535 |
| Base | 2107 | 11x | 23177 |
| Bull | 2408 | 14x | 33712 |
Earnings recovery bridge. Q2 PPOP ₹970.2 crore versus ₹694.8 crore prior year gives ₹275.4 crore improvement. Impairment fell ₹226.5 crore. Of ₹501.7 crore PBT uplift roughly 55% came from PPOP and 45% from lower impairment.
This is a stronger basis than a simple quadrupling of PAT: the franchise has actual operating expansion, but credit normalization was a major contributor. Sequentially PPOP +₹97.7 crore, loss +₹86.7 crore, so ~89% of incremental operating earnings was absorbed by higher impairment.
The quarter's PAT growth 1.7% therefore does not justify extrapolating the prior-year recovery rate. H1 PAT ₹995 crore annualizes ₹1,990 crore; Q2 annualized ₹2,006 crore is broadly similar. The next test is whether lower new PAR and operating income improve *both* PPOP and loss expense without relying on a lower tax or provision release.
| Bridge (₹cr) | YoY | QoQ |
|---|---|---|
| PPOP increment | 275.4 | 97.7 |
| Lower/(higher) impairment benefit | 226.5 | -86.7 |
| PBT increment (rounded) | 501.7 | 10.9 |
Unsecured mix is not automatically safer after graduation. Retail finance ₹7,200 crore, 23.3% of AUM, grew from 20.6% June. Secured retail only ₹695 crore (~9.7% retail, 2.2% group). Thus roughly ₹6,505 crore retail is not described as secured by these figures, and group JLG ₹23,746 crore.
A larger individual ticket can improve revenue and retention but increase borrower-specific risk and reduce group repayment discipline. The approximate retail AUM/borrower ₹1.06 lakh versus group proxy ₹62,200 suggests 70% larger exposure; borrower overlap and product definitions may affect the comparison.
Require a graduation cohort matched for prior loan history, household income, other lender count, ticket, tenor and region; compare net lifetime loss and repeat cash contribution. A borrower with strong group repayment may still be over-levered after moving to a larger individual loan.
The relevant outcome is not portfolio mix, but borrower-level economic return and affordability.
Loss normalization and reserve flow. H1 impairment ₹511.7 crore, write-offs ₹614 crore and management macro buffer ₹137 crore are different flows/stocks. Write-offs may exceed current charge because existing provisions absorb past losses.
They do not demonstrate cure. The reported Q2 nonannualized cost 0.99% roughly annualizes 3.96% on a stable denominator, near the upper end of 3–4% guidance; H1 annualized 3.43% includes 0.45% macro buffer.
The ₹299 crore Q2 impairment is higher than Q1 ₹212.5 crore. A useful downside model starts from annualized PPOP ₹3,881 crore and uses illustrative annual impairment 3%, 4% and 6% of closing AUM ₹30,946 crore.
Gross charges ~₹928/1,238/1,857 crore; approximate PBT ₹2,953/2,643/2,024 crore, PAT at assumed 25% tax ~₹2,215/1,982/1,518 crore. These are scenario sensitivities, not actual issuer average-AUM credit costs or forecasts; PPOP held constant, no change in income/expense/asset mix.
| Annual credit-loss sensitivity | 3% | 4% | 6% |
|---|---|---|---|
| Charge ₹cr | 928 | 1,238 | 1,857 |
| PBT ₹cr | 2,953 | 2,643 | 2,024 |
| PAT at 25% tax ₹cr | 2,215 | 1,982 | 1,518 |
The bear outcome is near the prior report's ₹1,505 crore normalized bear, demonstrating how modest deterioration in a high-yield lender can meaningfully reduce equity earnings. Actual loss on average assets, ECL methodology and tax differ.
The test is sustained low *new* PAR and cash cures rather than only the GNPA stock after write-offs. Geography-specific floods/drought can affect group cash flow together; borrowers may depend on the same local agricultural, trading and wage cycles even if lender limits appear granular.
Pricing and margin. Portfolio yield 22.3%, funding cost 9.1%, NIM 15.1%, guided 14.4–14.7%. NIM includes equity funding and accounting definitions; do not subtract funding cost from yield and call the result NIM.
A 50bp annual yield shock on closing managed AUM ~₹155 crore pretax. At 4% assumed normalized loss, combined sensitivity would reduce PBT from ₹2,643 to ~₹2,488 crore and PAT to ~₹1,866 crore, holding everything else static.
This explicitly challenges the assumption that the recent NIM peak persists. A rate cut can strengthen affordability and retention and reduce future losses; assess price elasticity and realized loss after pricing.
If margin drops but credit and customer retention improve, the economic return may be better. If both margin and loss deteriorate, valuation should fall.
Capital productivity and cash. September equity ₹8,919 crore and Q2 annualized PAT ₹2,006 crore imply ~22.5% closing-book return proxy, consistent directionally with management's 21–24% ROE guide but not the same denominator. A 14% illustrative equity charge ~₹1,249 crore; normalized PAT from the three loss cases creates residual ~₹966/733/269 crore on static book.
Actual equity growth and cost change these amounts. Sustained high return depends on cash conversion and retaining sufficient capital to grow a credit-intensive book. October signed ECB is not September liquidity; FX hedging and tenor determine actual cost.
The balance-sheet liquidity assets should be tested for encumbrance and drawdown risk.
Investment rule. Favor restoration of group loan growth accompanied by lower absolute new PAR, retail graduation cohort performance and stable margin after pricing changes. Compare actual credit charge with 3–4% guidance on matched periods, not a nonannualized quarterly rate.
Watch state cohorts, repeated borrowers and actual household indebtedness. A steady-state P/B at normalized ROE 15%/21%/24%, equity hurdle 16%/14%/13%, growth 5%/6%/7% is ~0.91x/1.88x/2.83x; this is terminal economics, not a price target, and cannot replace current audited book/share.
Strong reported PAT is insufficient if the next cohort creates another cycle of write-offs. Require evidence that the portfolio transition generates higher borrower-lifetime contribution, not merely higher ticket and retail mix.
The scale of the borrower franchise permits a more concrete view of new-customer economics. Of approximately 2.3 lakh new borrowers in Q2, the issuer reports 35% new-to-credit. On those rounded figures that is about 80,500 customers; the remainder, roughly 149,500, have some prior credit history.
This split cannot establish the proportion with outstanding loans at other institutions, because credit history and live indebtedness are different. It does, however, make two underwriting populations explicit.
New-to-credit applicants have thinner bureau records, while previously financed applicants may have multiple obligations or repayment histories that are not visible from one institution's collection record. The lending decision should reconcile household cash inflow, existing instalments, regional seasonality and product suitability for both groups.
A higher acquisition rate is valuable only when cohort-level lifetime cash contribution is positive after collections and losses.
Management's group-loan and retail-finance balances also allow a static mix risk budget. If incremental annual loss were 100 basis points on the ₹23,746 crore group book and 150 basis points on the ₹7,200 crore retail book, the resulting annual pretax charge would be approximately ₹237.5 crore plus ₹108 crore, or ₹345.5 crore.
These are hypothetical product-specific additional losses on closing exposure, not a forecast, and must not be added to another stress on the entire AUM. The asymmetric shock deliberately tests larger retail tickets rather than assuming a uniform loss change.
At an illustrative 25% tax rate the annual PAT impact would be roughly ₹259 crore if fully recognised in the same period, before funding and pricing offsets. That is about 12.9% of the Q2 annualised PAT run-rate.
The sensitivity is particularly useful because unsecured group and individual lending can be exposed to common household cash-flow shocks despite different collection structures.
A quarterly credit budget is another disciplined check against the 3–4% annual credit-cost guidance. On the September managed AUM proxy, a 3% annual rate implies roughly ₹232 crore per quarter and 4% roughly ₹309 crore, assuming a constant base and even quarterly recognition.
Actual issuer cost uses the defined average exposure and provisions can be uneven. Q2 impairment of ₹299.2 crore is close to the upper end of that rough range. At unchanged Q2 PPOP of ₹970.2 crore, those illustrative quarterly charges leave PBT of ₹738 crore or ₹661 crore respectively, before any other differences.
An investor should not assume that the reported quarter's strong profit can be maintained with 3% loss simply because management's range starts there. A regional stress episode, provision recalibration or new-cohort mix can change the timing and magnitude of expense.
The evidence needed is a reconciled stage roll-forward with cash collections from overdue accounts separated from write-offs and provision releases.
The October lending-price reduction should be evaluated through a borrower-payment model, not just the margin impact. A lower rate can reduce the instalment burden on qualifying new loans and improve borrower retention, but repricing existing loans depends on actual product terms.
With the reported 88% H1 retention, the remaining 12% is not automatically churn that can be recovered by a rate cut. Some borrowers complete their borrowing cycle, migrate or fail eligibility checks. The practical test is renewal rate among eligible borrowers with unchanged risk, their realised net yield and the subsequent delinquency of the renewed cohort.
This links competitive pricing to the economic objective of sustainable risk-adjusted borrower value.
Key takeaways
A positive catalyst is stronger H2 borrower acquisition and group recovery with stable X-bucket collection. A falsifier is new PAR increasing alongside write-offs, or larger individual loans to graduated borrowers producing worse loss rates. This report has original issuer presentation and call evidence, but full vintage curves, borrower overlap/bureau data, geography-level recovery and statutory share/ALM reconciliation remain unavailable. No formal price target or full-QA status is asserted. [1,2]
[1] Q2 FY27 investor presentation, 2026-10-22, original filing.
[2] Q2 FY27 earnings call, 2026-10-22, original filing.
Prices: Sage PRO canonical close, 13 November 2026. Analyst calculations and stress assumptions are explicitly labeled; no independent full-QA approval. Monetary figures are ₹ crore unless otherwise noted. Open issues: Diluted share count, state/product cohort losses, borrower overlap, independent ALM and peer comparisons.
https://www.bseindia.com/xml-data/corpfiling/AttachHis/9d90680b-2fb2-43a4-8818-9ec848ab6353.pdf
https://www.bseindia.com/xml-data/corpfiling/AttachHis/9d90680b-2fb2-43a4-8818-9ec848ab6353.pdf
Analyst scenario calculations
Original Q2 source ledger and analyst calculations
This is dated research, not a live market quote. Scenarios and historical performance are not guarantees.