ESAF Small Finance Bank - Bank That Just Came Back from the Edge

Bond Opportunities
1

I was reading about ESAF Small Finance Bank recently, and what stood out was how close it came to a difficult phase before rebuilding itself. ESAF Small Finance Bank has been issuing Tier II bonds at 11.30% coupon (Basel II compliant, unsecured, subordinated NCDs) through private placement. The most recent tranche was ₹150 crore in November 2025, with maturity in August 2031 (roughly 69 months). There was also a ₹50 crore tranche in August 2025. All part of a ₹1,000 crore capital raising plan approved by shareholders.

This is one of those names where the yield tells you half the story before you even open the financials. 11.3% from a scheduled commercial bank. That’s not normal. Let me walk through why.

Background first. ESAF SFB is a Kerala-headquartered small finance bank, promoted by K Paul Thomas who founded the ESAF group in 1992. It started as Evangelical Social Action Forum, an NGO doing microfinance. Got the SFB licence from RBI in 2015, started banking operations in March 2017. Listed on exchanges in November 2023 through an IPO that raised ₹463 crore. Currently operates 788 banking outlets, around 720 ATMs, and 1,042 customer service centres across 24 states and 2 union territories. Customer base of nearly 1 crore (99.9 lakh as of December 2025).

Now, the numbers. And this is where it gets uncomfortable.

FY24 (the good year):

  • PAT: ₹426 crore

  • NIM: 9.9%

  • GNPA: 4.76%

  • NNPA: 2.3%

  • CRAR: 23.3%

FY25 (the disaster):

  • Net loss: ₹521 crore. That’s not a typo. From +₹426 crore profit to -₹521 crore loss in one year.

  • NIM compressed to 8.1%

  • GNPA jumped to 6.87% (from 4.76%)

  • NNPA: 2.9%

  • Slippages: 10.29% (from 6.5% in FY24)

  • CRAR: 21.8%

FY26 trajectory (things got worse before they got better):

  • Q1 FY26: Loss of ₹81 crore, GNPA spiked to 7.48%

  • Q2 FY26: Loss of ₹116 crore, GNPA hit 8.5%, NNPA at 3.8%

  • Q3 FY26: Profit of ₹7 crore (first profitable quarter in a year), GNPA improved sharply to 5.6%, NNPA to 2.7%

So what happened? In short, the entire Indian microfinance sector blew up. ESAF’s loan book was historically dominated by unsecured micro loans (was 81% of the book in 2022). When the sector-wide stress hit starting Q2 FY25, ESAF was massively exposed. Slippages shot through the roof. The bank kept setting aside provisions (₹234 crore in Q1 FY26, ₹249 crore in Q2, ₹243 crore in Q3) and reported losses for five consecutive quarters.

The rating agencies responded accordingly. CARE Ratings revised outlook from Stable to Negative and downgraded subordinated bonds citing sustained weakening in asset quality and four consecutive quarterly losses. Brickwork Ratings downgraded Tier II bonds to BWR BBB+/Stable in August 2025, citing GNPA crossing 7% and five consecutive quarters of losses.

Now, the Q3 FY26 turnaround. The bank is telling a story of strategic pivot through what they call the MARG strategy (MSME, Agri, Retail, Gold loans). The shift is real in the numbers:

  • Secured assets now 63% of gross advances (was 45% a year ago)

  • Gold loans grew 40% YoY

  • Micro loan share has come down from 81% (2022) to roughly 37%

  • Disbursements surged 134% YoY in Q3 FY26

  • Gross advances grew 13.1% to ₹20,679 crore

  • Deposits up 7.1% to ₹24,006 crore

  • Retail deposits at 93% of total. That’s genuinely strong.

  • CASA ratio improved to 25.1%

  • NIM improved to 6.6% (from 5.9% in Q2 FY26)

  • Pre-provisioning operating profit jumped 171% sequentially to ₹253 crore

  • CRAR at 22.7%, comfortably above regulatory requirements

And importantly, GNPA dropped from 8.5% in Q2 to 5.6% in Q3. Part of this is organic improvement, part is ₹1,018 crore of NPA sales to ARCs in 9M FY26. They sold ₹733 crore of bad loans for ₹73 crore in June 2025 alone. So the balance sheet is being actively cleaned up, though the recoveries on ARC sales tell you how bad some of that portfolio was. Roughly 10 paise on the rupee.

Management guidance from the Q3 earnings call (February 2, 2026):

  • Credit costs expected to normalise by Q1 FY27

  • Targeting steady-state ROA of 1.5-2% by FY28

  • Loan book growth ~15-16% for FY26, ~25% for FY27

  • Planning to cap unsecured book at 30-35% long-term

  • Digital transformation (ESAF 2.0 StratoNeXt) targeted for Q2 FY27

  • Universal banking licence application being explored but “at least two years away” per the MD

The real question for bondholders: Is 11.3% adequate compensation for the risk you’re taking here?

On one hand, ESAF is a regulated scheduled commercial bank with CRAR above 22%, the promoter has stated willingness to infuse capital, retail deposits fund 93% of the book, there’s no ALM mismatch crisis, and the turnaround is showing up in the numbers.

On the other hand, these are Tier II subordinated bonds. In a winding-up scenario, you get paid after depositors and senior creditors. The rating is A/Negative from CARE and BBB+ from Brickwork, which means one more downgrade from Brickwork puts you at BBB territory. The bank lost ₹521 crore in FY25. Five consecutive quarterly losses eroded net worth. And the RBI recently rejected the promoter entity’s (Dia Vikas Capital) plan to acquire shares exceeding 5%, which adds corporate governance uncertainty.

The microfinance sector stress isn’t fully behind us either. The improvement in Q3 is real but one good quarter doesn’t make a trend. Slippages were 10.3% in FY25, and even if they halve in FY26, that’s still elevated.

Geographic concentration is another factor. Kerala, Tamil Nadu and Madhya Pradesh together account for 67.6% of the portfolio. Kerala alone is 34.8%. Any regional stress disproportionately hits ESAF.

Disc: Not invested. Studying this carefully before taking a position.

2

I’ll be direct. The Tier II subordination is the thing that should give everyone pause, regardless of the yield.

Let me explain why this matters more for ESAF than for, say, an SBI or HDFC Bank Tier II bond. In a well-capitalised, consistently profitable bank, the subordination is largely theoretical. You’ll never actually face a loss absorption scenario because the bank is nowhere near resolution territory.

ESAF is different. This is a bank that lost ₹521 crore in one year. Net worth as of December 2025 is ₹1,756 crore. If FY25-level losses continued for another 3-4 years (not predicting this, but it’s within the realm of possibility if MFI stress returns) the capital buffer starts looking thin. And Basel II Tier II bonds have a specific feature: they can be written down if the bank reaches a point of non-viability as determined by RBI. That’s not a default clause. That’s a regulatory write-down clause. Your principal can go to zero without the bank technically “defaulting” in the traditional sense.

Now, do I think ESAF is headed for non-viability? No, probably not. CRAR at 22.7% gives reasonable buffer. The promoter has stated capital infusion intent. The MARG pivot is directionally right. But when you’re buying subordinated debt from a loss-making bank with deteriorating asset quality and a negative rating outlook, the tail risk has to be priced in.

At 11.3%, I don’t think the risk-reward is compelling enough for that tail risk. I’d want 13-14% to compensate for the subordination plus the asset quality uncertainty plus the potential further downgrade. The fact that secondary yields are already at ~12% suggests the market agrees.

If ESAF were issuing senior secured NCDs at 11.3%, that would be a very different conversation.

3

Want to push back a little on Anshul’s framing. Yes, the tail risk is real, but let me put some numbers around the recovery story because I think it’s being under-appreciated.

The Q3 FY26 numbers aren’t just “one good quarter.” The sequential improvement across every metric is striking:

GNPA: 7.48% → 8.5% → 5.6% (Q1 → Q2 → Q3) NNPA: 3.8% → 3.8% → 2.7% PPOP: ₹125 Cr → ₹93 Cr → ₹253 Cr NIM: 6.0% → 5.9% → 6.6% Result: -₹81 Cr → -₹116 Cr → +₹7 Cr

The GNPA drop from 8.5% to 5.6% in one quarter is massive. Even adjusting for the ARC sales, the underlying collection efficiency is improving. Fresh slippages are declining. And the shift to 63% secured book means the portfolio generating these NPAs is actively shrinking as a proportion of the total.

I’m not saying ESAF is out of the woods. But consider the maturity: August 2031. That’s roughly 5.5 years out. By management’s own guidance, they expect to be at 1.5-2% ROA by FY28. If the MARG strategy holds and credit costs normalise by Q1 FY27 as guided, this bank looks very different 18 months from now.

Also worth noting, CRAR at 22.7% vs regulatory minimum of 15% for SFBs. That’s 770 basis points of buffer. Even if the bank continues to bleed losses for another 2-3 quarters (which the Q3 trend doesn’t suggest), the capital adequacy remains well above regulatory threshold.

The real question is whether you believe the MFI sector stress is largely behind us. If it is, ESAF at 11.3% with improving fundamentals is a genuinely attractive yield pick. If you think there’s a second wave of MFI stress coming, then even 14% might not be enough.

4

Good debate. Let me add a couple of practical points for people evaluating this.

First, the geographic concentration risk is underappreciated in this thread. Kerala at 34.8% of the book, top 3 states at 67.6%. For comparison, most of the NBFCs we discuss here have much broader geographic spread. When the Kerala floods hit in 2018 or when any state-level political event disrupts collections, ESAF takes a disproportionate hit. The bank has been reducing Kerala concentration (was 45.6% in March 2022), so the direction is right, but 34.8% in one state is still significant for a bondholder underwriting a 5+ year tenor.

Second, the RBI rejecting Dia Vikas Capital’s shareholding plan is a wildcard that nobody’s talking about. ESAF needs to comply with RBI Bank Ownership Directions 2023, and the promoter shareholding restructuring they had planned just got rejected. The board is “evaluating implications and exploring alternatives.” That’s corporate governance uncertainty that you’re not getting paid extra for.

Third, I want to flag the difference between what MarketsMojo / standalone quarterly numbers show (GNPA 8.54% in Q3) and what management presented in the earnings call (GNPA 5.6% in Q3). The 8.54% number appears to be based on exchange filings that may not reflect the ARC sale impact timing. The management-reported 5.6% post-ARC-sales is the more relevant number for credit analysis, but it’s worth being aware that the headline NPA picture depends heavily on when you measure and whether you adjust for sales. Without those ARC sales, the underlying GNPA would still be much higher.

Fourth, for anyone who hasn’t bought subordinated bank debt before. These instruments have a specific call feature and loss absorption mechanism that regular corporate NCDs don’t have. Make sure you understand the Basel II/III Tier II documentation before investing. The Brickwork rationale specifically notes the downgrade was driven by GNPA crossing 7% and continued losses impacting ROA. If those triggers persist, further downgrades are mechanical.

5

Something that isn’t getting enough attention: ESAF sold ₹733 crore of NPAs for ₹73 crore in June 2025. That’s a recovery rate of about 10%. Then another ₹362 crore to ARCs in Q1 FY26. Total ARC sales of ₹1,018 crore in the first nine months of FY26.

That tells you two things. One, the NPA improvement story is partly real collection improvement and partly accounting. Transferring bad loans to ARCs at steep discounts cleans up the GNPA ratio but you’re crystallising losses. Those ₹733 crore of loans that went for ₹73 crore? That’s ₹660 crore of value destruction that’s already happened. It’s priced in, but it tells you how bad the book was.

Two, and this is the more interesting point, it tells you management is being proactive about cleaning up the balance sheet rather than letting NPAs linger. I’d much rather see a bank sell ₹733 crore at 10 paise and move on than carry ₹733 crore of deadweight on the books while pretending collections will improve. The Keertana thread had a similar discussion about the difference between NBFCs that recognise stress early vs those that hide it. ESAF seems to be in the “recognise and move on” camp, which is actually a positive from a credit standpoint even though the numbers look ugly.

The question is: once you’ve cleaned out the legacy MFI stress and shifted to 63% secured, what does the steady-state credit cost look like? Management says 2-3% for the current portfolio mix. If that holds, and NIMs stay at 6.5%+, you can build a path back to profitability that’s sustainable. But it requires the new secured book (gold, LAP, MSME, agri) to perform well, and that book doesn’t have enough seasoning yet to be confident about.

6

Good points from everyone. Let me try to distil this into a practical framework.

The bull case: ESAF is a regulated bank with strong CRAR (22.7%), turning the corner on asset quality, actively de-risking by shifting to secured lending, and the 11.3% yield compensates you for the near-term noise. The MFI cycle is arguably peaking, Q3 FY26 shows genuine improvement, and you’re getting paid 500+ bps over comparable SFB FD rates to hold a 5.5-year bond from a bank that’s likely to look much healthier by maturity.

The bear case: You’re buying subordinated debt from a bank that just lost ₹521 crore, has been downgraded by two rating agencies, has a negative outlook from CARE, is cleaning up its book via ARC sales at 10 paise recovery, has an unresolved promoter shareholding issue with RBI, and is concentrated in 3 states. The yield is 11.3% but secondary is already at ~12%, suggesting the market thinks the risk premium should be higher. One more bad quarter or a second wave of MFI stress and you could see another downgrade, at which point the secondary value of this paper drops materially.

My view: This isn’t a name I’d put more than 2-3% of a fixed income portfolio into, and only if you genuinely believe the MFI cycle has peaked. The CRAR buffer and the pivot to secured lending are real positives. But the subordination risk, negative rating outlook, and promoter governance issues mean this is firmly in the “opportunistic high-yield allocation” bucket, not a core holding.

I’d want to see at least two more quarters of profitability and GNPA staying below 5% before getting comfortable. If the bank delivers that through Q1 FY27, I think the CARE outlook flips to Stable and the pricing becomes more interesting. Right now, the risk-reward is closer to fair than compelling.

Will update when Q4 FY26 results come out and if there’s any CARE rating action post the Q3 improvement.

Disc: Not invested. Monitoring.