CreditAccess Grameen Ltd

MR. RENISH BHUVA - ICICI SECURITIES LIMITED

Page 2 of 19 Ladies and gentlemen, good day, and welcome to CreditAccess Grameen Limited Q2 FY25 Earnings Conference Call hosted by ICICI Securities Limited.

As a reminder, all participant lines will be in listen-only mode and there will be an opportunity for you to ask questions after the presentation concludes. Should you need assistance during this conference call, please signal an operator by pressing ‘*’ then ‘0’ on your touch-tone phone.

Please note that this conference is being recorded.

I am now handing over the conference to Mr. Renish Bhuva from ICICI Securities. Thank you, and over to you, sir.

Hi, thanks Nirav. Good evening, everyone, and welcome to CreditAccess Q2 FY25 Earnings Call. On behalf of ICICI Securities, I would like to thank the CA Grameen Management Team for giving us the opportunity to host this call.

Today we have with us the entire top management team of CA Grameen represented by Mr.

Udaya Kumar Hebbar - Managing Director; Mr. Ganesh Narayanan - CEO; Mr. Nilesh Dalvi - CFO and Mr. Sahib Sharma - DGM, Investor Relations.

I will now hand over the call to Mr. Udaya Kumar Hebbar for his opening remarks, and then we will open the floor for Q&A. Over to you, sir.

Thank you, Renish. Good evening, everyone and thank you for joining the conference call to discuss our second quarter and the first half of FY25 business performance. The unsecured lending space, marked by high growth, has been witnessing an increase in delinquency levels over the past 3-4 quarters. Given the short-term nature of unsecured lending and timely calibration by various institutions, we believe this credit cycle to be transient in nature. Taking due cognizance of this, the microfinance industry represented by MFIN, implemented the guardrails in July-2024 to strengthen the underwriting norms across all regulated entities. While we have been witnessing an accelerated realisation of delinquencies in the last few months, we believe that the MFIN guardrails are essential for fostering a healthier environment for the industry in the medium to long term.

Coming to the second quarter performance, the overall AUM grew by 11.8% YoY to INR 25,133 Crore with the GL book growing by 9.3% YoY to INR 24,188 Crore. Meanwhile, the RF book continued to show robust growth driven by selective customer base, with an AUM of INR 945 Crore. The customer base grew 7.2% YoY to 49.33 Lakh at the end of Sep-24. On a QoQ basis, we witnessed 4.4% decline in the overall AUM and 1.0% decline in customer base. We added 1.46 Lakh new customers during the quarter and 3.37 Lakh during H1 FY25. Our branch infrastructure expanded to 2,031 across 398 districts as we added 55 new branches during the quarter.

Page 3 of 19 The Net interest income grew by 20.8% YoY to INR 933 Crore. As guided before, our average and marginal cost of borrowings remained steady at 9.8% and 9.4% respectively. Our portfolio yield at 21.1% and interest spread of 11.4% too remained stable and continued to be one of the lowest in the microfinance industry. NIM stood at 13.5% for Q2 FY25 compared to 13.0% in Q1 FY25, because of higher capital adequacy of 26.2% and low base effect on account of a decline in loan portfolio. The cost-to-income ratio was 30.7% while PPOP grew by 19.5% YoY to INR 672 Crore at the end of Q2 FY25.

We would now like to draw your attention to slides 6 to 10 on the asset quality update. We have observed a temporary increase in delinquencies across various geographies on account of several factors. There have been localized disruption in repayments caused by third-party interventions in various states. Customers have been facing tight liquidity and cash flow constraints as the lending industry has curtailed disbursement with more focus on controlling delinquencies.

Further, certain customers, especially agri-laborers, have also faced income variations on account of low rainfall last year followed by severe heat waves during April to June '24.

Furthermore, there has been also a transient impact on heavy rainfall in several regions during September quarter.

On top of all the above, a segment of over-leveraged borrowers with lower cash flow also are part of this delinquent bucket. We expect the situation to stabilize in Q3 FY25 and business sentiment should improve in Q4 FY25. The company continues to strengthen its collection effort with additional PAR control measures by deploying senior and experienced field staff and quality control teams on the back of continuous customer engagement in addition to strong underwriting measures.

We have tried to understand the current delinquency trend by analyzing the customer overlap with other lenders along with leverage and vintage. The delinquency is the lowest in the case of unique customers with high vintage and highest in the case of customers with low vintage and high velocity of loans being availed from multiple lenders. As on Sep-24, [CA Grameen + 4 & above] accounted for 12.6% of the group lending portfolio and had PAR 15+ of 12.2%. Hence, the overall PAR 15+ impact on account of [CA Grameen + 4 & above] customers is only 1.5%.

We have also drawn state-wise comparisons of delinquency versus the microfinance industry by sourcing Credit Bureau data for the month of Aug-24. As visible on slide 9, our performance has been relatively better than the industry across our top five states, accounting for over 85% of our gross loan portfolio.

Our collection efficiency (excl. arrears) stood at 96.3% for Q2 FY25, PAR 90+ at 1.74%, GNPA of 2.44%, and NNPA of 0.76% both predominantly measured at 60+ DPD. The credit cost stood at INR 420 Crore for Q2 FY25 which includes an additional INR 29.8 Crore provision on account of increased ECL% across all three stages compared to Q1 FY25. Overall, the annualized gross credit cost stood at 4.7% for H1 FY25.

Page 4 of 19 While analyzing our credit cost figure, we would like to draw your attention to our early stress recognition and conservative provisioning policy. The Company is currently holding ~179 bps (INR 431.1 Crore) higher provisioning over PAR 90+ and ~256 bps (INR 629.2 Crore) higher provisioning compared to IRAC prudential norms. Even when we compare us with the average provisioning policy adopted by the NBFC industry, we are currently holding additional INR 102 Crore provisions on account of our early recognition and higher provisioning rates. This will help us to recognize 70-75% of the current stress in the current financial year instead of deferring it to the next year.

We had a comfortable liquidity position as on Sep-24 with cash and cash equivalents of INR 2,036 Crore amounting to 7.6% of total assets. We have sanctions in hand of INR 3,830 Crore and another INR 6,918 Crore worth sanctions in the pipeline. Our capital adequacy remains strong at 26.1%, further adding strength to our balance sheet.

Our Q2 FY25 PAT stood at INR 186 Crore, with a ROA of 2.7% and ROE of 10.7%. For H1 FY25, the PAT stood at INR 584 Crore with a ROA of 4.1% and ROE of 17.1% reflecting our strong operating profitability on the back of resilient business model. In light of the current industry landscape and short-term challenges encountered, we have revised our estimate for FY25 annual performance guidance. We anticipate loan portfolio growth of 8-12%, NIM of 12.8-13.0%, credit cost of 4.5-5.0%, ROA of 3.0-3.5%, and ROE of 12.0-14.0%. We note that the revised guidance is based on our estimation of stabilization of delinquencies in Q3 FY25 followed by improved business momentum in Q4 FY25. This should be on the back of improved rural economic dynamics during the second half of FY25 driven by healthy agricultural activities. The country has witnessed the highest monsoon rainfall over the past 4 years which should help in increased sowing activities resulting in a positive spillover effect. We reiterate our medium-term growth outlook aiming to reach INR 50,000 Crore mark by FY28 as guided earlier through a combination of both Microfinance and Retail Finance businesses.

Lastly, we would like to share with you all an important update related to the management team.

Mr. Gururaj Rao has been elevated to Chief Operating Officer and KMP effective from 01st November, who was serving as the Chief Audit Officer. Also, Mr. Nagananda Kumar, currently serving as the Head – Operations, is now appointed as the Head – Internal Audit with effect from 01st November 2024. Both of them have been associated with us for more than 15 years, contributing to the Company’s growth and success. I am confident that under their leadership, we will further strengthen our market share and deliver exceptional value to all our stakeholders.

Thank you for your patient hearing. We look forward to addressing your queries as we open the forum for questions and answers.

Page 5 of 19 In the meantime, I am requesting Nilesh to give you a 5-minute brief on the asset quality slide, so that he will give a little more clarity on those slides.

Hi, good evening, everyone. So, I will be referring to Slide #6 to 10. Since there is a lot of data we have shared, I thought it is better we provide you some interpretation of the data given.

On Slide #6, we have given an update on how the PAR has moved up from Q1 to Q2. So, overall, PAR 0 has increased by 2.9% and PAR 90 has increased by around 1% and PAR 60 has increased by around 1.5%. These percentages are on a pre-write-off basis.

Overall, the slippage what we say at 60 DPD, is around 1.5% and at 90 DPD, the slippage is around close to 1%, the new PAR accretion. And we have also given the PAR0 figures across our top 5 states. So, Karnataka, it is 2.3%, Maharashtra and Madhya Pradesh relatively they are holding well at 4.2% and 4.6%. In TN, it has increased to 6.2%, and Bihar is on the higher side at 8.9%. And other states, it is 8%, but here we have to also take into account that the book de- growth is higher because there are certain states where we are kind of not growing much like Rajasthan, Gujarat, Kerala, etc. So, to that extent, there is a base effect which is kind of playing out there.

Now we move to the Slide #7. Here we have just tried to analyze the PAR by looking at the vintage, lender overlap and leverage. So, the first chart shows the portfolio breakup. This is the lender overlap and the vintage. We see that unique customers, they are accounting for around 27% of the book in AUM terms and CA Grameen +4 and above that segment is accounting to around 12.6% of the overall portfolio. Here we see that the larger proportion of customers in 4 and above bucket, it is in the low vintage, that is between 0 to 4 years. And the table to the bottom right, there we have given the PAR 15+, that is Stage-2 and Stage-3 across lender overlap and the vintage.

So, again, here we see that the PAR is lower in the unique and CA Grameen +1, +2 categories, whereas it inches up in the +3 and +4 and above buckets. CA Grameen +4 and above it is around 12.2% in the PAR 15 as on Sep-24 end. Again, here we see that the PAR is higher in the 0 to 4 years of vintage bucket.

Largely, what we have inferred from this is that customers who we recently onboarded over last couple of years who have relatively lower vintage with us, that is where since we cannot meet their entire requirement because we will start with say INR 40,000-45,000 kind of a loan, that is where they have larger overlap with other lenders because maybe they might have a certain history in the microfinance industry and their overall funding requirement is higher.

I wanted to highlight is the CA Grameen +4 and above cohort, it is 12.6% of the overall group loan portfolio where the PAR 15+ is 12.2%. So, overall, CA Grameen +4 and above cohort

Page 6 of 19 accounts for 1.5% PAR on the overall book. Out of whatever 5% PAR we have, out of that the impact of this cohort is around 1.5%, CA Grameen +4 and above cohort.

Now, moving to Slide #8, here we have again tried to analyze the PAR by considering the leverage. Here, as we see in the first table to the top left, the total indebtedness of more than INR 2 Lakh, there it is around 9.4%, the overall borrowers, and the PAR 15 for that cohort, that is more than INR 2 Lakh, it is 6%. So, again here, as we see, the PAR is lower for high vintage customers even though their overall indebtedness is higher. As we see in the table in the bottom left and the PAR is higher in the case of low vintage customers who have taken higher leverage.

Again, the point we are trying to reiterate is that if higher leverage is aligned with vintage, then the PAR is lower and only in early vintages, if customers have taken multiple loans or they have higher leverage, that is where the PAR is relatively higher. Now considering the two tables on the right-hand side, wherein we have tried to understand CA Grameen +4 and above borrowers, 4 and above borrowers with more than 2 Lakh leverage, that is 6% of the overall borrower base, and there the PAR 15 is around 10%. Again in the bottom right table, the same point appears again that the PAR is higher in the low vintage buckets where the leverage is higher and as the vintage increases, the PAR is relatively lower.

Now, we go to the next slide, that is Slide #9. Here we have tried to assess our performance versus industry during March to August. So, across all our operating states, PAR 1 to 180, so we are considering 1 to 180 because the write-off policy varies from player to player across industry.

That is where the 180 plus, there is a large variance. That is why we are considering 1 to 180 bucket and that is where as on August-24 end, CA Grameen's PAR 1 to 180 was 3.4% versus 6.0% for industry.

In our top four states, we are doing better than the industry. In the case of Bihar and UP, we are largely in line with the industry. So overall, the top five states which account for 85% of our portfolio, there we are relatively doing much better than the industry. This is the CRIF Highmark data what we have sourced for the month of August, which is the latest available data.

Moving to the next slide, on Slide #10, we have indicated the ECL what we have taken as on Sep-24. So here, the thing is that since ECL is a forward-looking model, we keep revising the underlying data every quarter. We typically take 36 data points, three-year data. And as the recent data enters into the computation, the ECL rates vary.

So here, from Q1 to Q2, we have seen the ECL rates have increased across Stage-1 and Stage- 2. In Stage 1 last quarter, our ECL was 0.89%, which is now 1.0% and in Stage-2, it has gone to 57.6% from 56.7%. So, because of this, it is helping us to build early provisioning buffers and that is where today we see that compared to IRAC norms, we are almost holding over INR 600 Crore of additional provisions, and even when we compare it with the NBFC industry norms, wherein the Stage-2 is provided after 30 days and Stage-3 is provided after 90 days, since we do

Page 7 of 19 early recognition, we are providing INR 100 Crore of additional provision by booking Stage-2 after 15 and Stage-3 after 60.

To that extent, we believe that the current delinquency which has arisen, almost 60% of the impact has been taken in the P&L and that gives us confidence that the current delinquencies will be largely taken care in this year, so that the spillover effect in the next year will be relatively limited.

These are the larger points I wanted to make. We can now take questions.

Thank you very much. We will now begin with the question and answer session. The first question is from line of Dhaval from DSP Mutual Fund. Please go ahead.

Thanks for sharing all these additional data points that are quite useful. I have two questions.

First is related to Slide #6 and Slide #9. So, if you look at the change in August to September in PAR, it seems quite sharp in certain states. The PAR increase has been quite sharp. Is that how you expect the, so, the first part is, when do you think this number starts stabilizing based on the past experience? And so, that is one part of the question. And the second is, given that you said 60% of the recognition or at least the provision part is taken, do you think this is the peak level of credit cost and we should start seeing moderation from next quarter?

First question, PAR, it is not August to September, it is June to September, I think, what you asked. The change is from June to September and is not in the slide 6.

Sorry, sir, what I was referring to, so if you look at the state-level PAR that was shared, CREDAG versus industry, I think, in slide 9, where we have given, for example, states like Bihar and MP and Karnataka, etc., so, that PAR is as of August for the industry and for us as well.

And in September, we have given PAR data in, I think, slide 6. So, the delta change that I am seeing in each of the state numbers is quite sharp. I just wanted to be concerned that that is a kind of magnitude of increase that one has seen and when we think this will stabilize, especially in the top five states.

I think there are some data variations. So PAR 90, it is showing. See, Dhaval, that is the difference. That is 180, this is 1 to 270. That is why there is a difference between that because we don't have 1 to 270 proper data from the system, from the Highmark. That is why we took comparable data which is 1 to 180 from the Highmark data. So, our data is 1 to 270 days.

Therefore, there will be some differences.

The underlying question is basically, this PAR increase that you would have seen, PAR 0 is what I am referring to, between August and September, what would be the kind of change that you would have seen in the top five states? And when do you think this starts peaking out? That's the underlying question actually.

Page 8 of 19 Actually, there may be a difference between 75 bps to 100 bps. That's the difference between that. Now every month it's a variation of 50 bps to 75 bps. So, maybe September was a kind of peak. It could be above. We don't have the right exact data here because quarter-to-quarter data is available with me here, but it could be a little higher about 75 bps, it could be higher than August to September. Just from estimation, it may not be too different.

And on the second part, you said about whether this is the peak. So, our view also should be peaked, but credit cost wise it may not be. What I am telling is the PAR accretion or increase should be peak as per us this quarter or maybe the trend of reversal should start from Q3. So, it may be October may be stabilized, November may start turning around, and December may be bit improvement. So that kind of change we are estimating for Q3.

And so credit cost, you are saying that should peak in Q3, so credit cost wise? Yes, that is our estimate.

And the bulk of the pain should be, I mean 80%-90% of the pain should be crystallized in FY25. Is that how one should think about it?

That is exactly what we are talking about because our 15+ is carrying almost 60% average, right, already. Every quarter, we are already carrying more than 60% for the 15+DPD book. So, by default even at the March level, we would have carried 60% plus for the 15+DPD already, right.

So, automatically, we will be carrying, we will be crystallizing majority of the cost in the same financial year.

And just one last question in terms of growth, how should one think about this current environment and going forward for FY26 as well, from a growth standpoint, if you have any thoughts?

Our view is that we should see a turnaround from Q4 and this continue afterwards. I think we don't see any challenge in FY25-26. So, that is why we are reiterating the medium-term growth or medium-term, what you call, guidance for INR 50,000 Crores by FY28. We are reiterating that.

Next question is from the line of Rajiv Mehta from Yes Securities. Please go ahead.

Sir, what was the regular bucket collection efficiency in Q2 and versus Q1? And is October showing an improvement in that parameter?

So, Q2 was 96% and Q1, was 97% actually. That is a 1% variation.

I am asking for the regular bucket, 0 DPD bucket collection efficiency, specifically you have that number?

Page 9 of 19 Yes, Rajiv. So, the on-time collection is 95% for September. Last quarter it was 96%. And October?

As on September end it's 95%. So, usually, you can take it as whatever is in PAR, that will be the non-standard. So, it should be 95%.

And then is it the right thing to understand that this metric has to start stabilizing and improving for at least the fresh PAR creation will have to start stopping for the whole situation to stabilize?

Is that the right way to monitor this number?

No, it will not stop, but it will reduce from here on. That is our estimation. We believe that September may be the peak, with October stabilizing at the same level before starting to decline.

This is our estimation, and we are already observing this trend in a couple of cases.

So, what is the parameter in terms of collection metric which you are looking at which will make you start disbursing again more assertively?

No, we are actually dividing branches between the stable branch and a bit of non-stable branch.

In the non-stable branch and stable centers, we are planning to start disbursing. We are already disbursing to some extent, and we will start to look more carefully in this upcoming quarter, specifically in Q3.

So, our metric is more of a stable branch and stable center for growth and the non-stable branch and non-stable center is more of a focus on control and collections. From a data perspective, I will tell you, 70% of centers have a 0 DPD actually. So, that's why we are actually working on how to renew well, how to renew with better underwriting and all. 73% for the sake of numbers, 73% of our customer centers have 0 DPD.

And sir, your 0 to 60 DPD bucket has fallen from 1.1% as of June to 2.5% as of September. I am sure you have put a lot of collection effort there, but what is your assessment? Would it largely flow into 60+, or can you reduce the flow because of whatever effort you have put?

Unfortunately, our view is the flow is higher, not lower, because these are the type of customers, as I said earlier also, it is more customers we acquired in the last 1 to 2 years, probably at the higher multiple lending segment. So, we are not seeing too much recovery from there. As I said, we are very careful and providing it upfront. It will come back to some extent, but not too much.

The flow is a little higher compared to earlier.

Page 10 of 19 And sir, one last question. When I look at PAR 0 and PAR 30 figures in Bihar, in the newer states, it is pretty high versus the vintage states, versus the older states, but even in Bihar, we are on a very similar line with the industry. There is no difference between CREDAG and other player or the industry as such. Generally, that is not something that we have seen in the vintage states.

Bihar is only an exception at this point of time.

Even in the newer states, I believe that when I look at even in the newer states for which you have not given the data of 1 to 180, I am sure you are very much around the industry there.

There we explained, where we reduce our portfolio drastically, there is an impact of the base effect also. At least 2% variation is because of the business impact, the denominator impact is there. Bihar and UP largely we are in line with industry.

The next question is from the line of Gaurav Kochar from Mirae Asset. Please go ahead.

Three questions from my side. Firstly, how much of this PAR 0 buildup that you have seen in September? How much of this would you attribute to, let's say, the late monsoon that we saw in floods in various areas? And you have also mentioned in one of the slides that there were floods and there were collection-related challenges. How much of that would you attribute? And in sync with that, let's say, in October, has that PAR improved?

So, flood-related is about INR 70-80 Crore. No, last September. Now it has actually already come back to almost INR 20 Crore. Normally, there is a small impact on this number because of flood related, but that will not significantly impact us. It will actually come back. It has already come back to a large extent. So, we should not worry about that. That is not a big part. It is a small part actually in the September numbers. When again October some more would have gone up in the last couple of days, but again those will come back. What was the second question?

Yes, just in sync with this, let's say, the overall October collection efficiency front, has that improved over September so far, while there are still five more days?

So far it has remained stable alongside September, Gaurav. There haven’t been many changes, that’s what I was saying. October and November will be more about stability and then we should look for improvements. Maybe in November, we will see that little improvement and in December we should see more improvements.

And sir, the AUM growth guidance that you given 8%-12% for the full year, now even if I take the lower end of that guidance 8%, that is translating into (+15%) growth in the second half of this year. Given the trends currently when we are in stabilization sort of a mode, are you comfortable to grow this portfolio by 15% over the September24 levels in the second half?

Page 11 of 19 Yes, that is our estimation. In Q3, we should be able to turn around. That is what the symptoms show in our book as well as in our field level estimation. If that happens, we will get almost four months of disbursement time, which is actually quite sufficient for growing, because we have a very large number of customers where renewals are coming in the second half.

Normally, in the microfinance business, the second half has higher renewals. And we have a very large number of retained customers. So, obviously we will be targeting our own renewal customers. So, we have a good chance of achieving it.

Just the last question, if I can squeeze in. Your credit cost guidance of 4.5% to 5.0% for the full year, if I look at the first half, you delivered, and if I annualize that, it is closer to 4.6% to 4.7%.

And you are saying that you expect the collection trends to stabilize in Q3 and in Q4 it should improve. So, is it more on the conservative side or do you believe that overall credit cost outcomes may be slightly better than that? Is it maybe more conservative on the guidance side?

Because if I do the math, it translates into, let's say around INR 700 or INR 720 odd Crore kind of numbers, which is pending in the second half. So, do you expect that sort of credit cost in the second half?

That is why we kept it a little larger, what you call range, right? So, the larger range we kept, because the best case could be the lower range, and the worst case should be the other range. So, that is why we kept that. So, because this little uncertainty is there everywhere, right? You can't be very precisive as of today.

That is why if you see our entire guidance has a larger range compared to our earlier guidances.

Because there is uncertainty, we have some plans, some indications. Accordingly, we are working. So, our view is we should do better. That is why if we do better, we will get a lower range.

The next question is from Abhijit Tibrewal from Motilal Oswal. Please go ahead.

Two things I want to understand. The first thing is in the slide and even in your opening commentary, when you call this a transitory or temporary increase in delinquencies, I think what we are wanting to communicate is that maybe this credit cycle that we have seen in the MFI sector today that could be more short-lived, could get over in this year rather than transitory in the sense that these are temporary operational difficulties and will get course corrected in the coming quarter. Is my kind of understanding correct when we use a word transient or temporary?

Yes, exactly, but it is not necessarily operational difficulty. It is a bit of environmental issues also there. Environment, I am talking about the whole industry sector. It is not that the micro- finance industry is struggling. There are a lot of other industries which are the finance sector, micro-finance sector, individual loan sector, personal loan sector, credit card sector, retail loan sector. You can see all of them maybe because of some of the reasons which are in the nature of

Page 12 of 19 third-party intervention, some of the nature of lack of income, some of the nature of the lack of liquidity and cash flow because of tight credit norms by everybody. All these are mixed here.

So, that is why we are saying because these are short-term loan cycles, so this cycle also should be short-term and should be over in this financial year itself. That is what our view.

Abhijit, one more point to add here. So, the thing is that the MFIN guardrails which have been introduced in the month of July, so that will or that has significantly tightened the underwriting norms in the industry. What we are seeing is that because of this, it will be more of accelerated recognition of the stress. What we have been seeing since last year September, there has been a steady-state increase in the delinquencies on account of various factors.

So, now it is more of accelerated recognition of these delinquencies which will happen in this financial year or which started in July, June and maybe it should end somewhere in the third quarter. So, to that extent, as you said rightly, it will be a short-lived delinquency cycle and post that, the industry delinquency trend should be relatively better.

Next question is from the line of Shweta from Elara Capital. Please go ahead.

So, two questions. One is, sir, you have implemented a district-based pricing model, but if I look at Q2 versus Q1 yields, they have remained largely stable. And the second question relates to you Nilesh. So, you mentioned that customers with 4 and above lenders are largely the low vintage portfolio customers and the overlap with other lenders is higher primarily because your origination amount or initial amount has been as low as INR 40,000 to INR 45,000. So, does it mean that these customers initially came to you and then suddenly they moved to the other lenders? That's why the number of lender count has increased because you were giving less amount of loans in the initial phase.

Yes, Shweta, I will take the second question first. So, the point I was trying to make is that, as you know, every year, whatever new customers we are adding, 30% to 35% of the customers are new to credit, which means that the balance of 65% to 70% customers are from the existing market and the unique proportion is higher in the core markets wherein we have a longer operating history. And in the newer markets, it's relatively lower around say 20% - 25% or 25% - 30%.

What happens is that when we go in newer markets and we are tapping customers from the industry, it is very much possible that the customer has been operating in the microfinance industry for 5 to 10 years. And since she is new to us, we have to start at level zero, wherein we will be lending anywhere between INR 40,000 to 50,000. So, that is where since we are not able to meet their entire requirement, they might end up availing loans from two or three lenders more.

That is the point I was trying to refer. And as the customers build higher vintage with us, their eligibility keeps increasing cycle after cycle, post which we are able to largely fulfill their entire financial requirements. That is where in the higher vintage customers we see that we have a higher wallet share.

Yes, on this District-based spacing, yes, we implemented it already. So, there is a bit of a shuffle between the districts from the state level to the district level now. But it should not change too much numbers because it only starts catching up quarter-by-quarter over a period of time, it may change a bit. But it will actually help us to identify the risks and place this properly at the district level, at a granular level. It will not change too much on the yield immediately, and it is actually not anticipated to have a yield change. It will remain stable because of this. We have actually degrown during the quarter. So, it will not make too much difference.

The next question is from the line of Nikhil Rungta from LIC Mutual Fund. Please go ahead.

Sir, just two questions, I will say it in one go. One is, we have seen unseasonal rains in October.

Of course, you have mentioned that in October, the collection efficiency should be similar or better than September, but because of this unseasonal rain, do we see any impact on October efficiency as well?

Second, when will disbursement growth come back. And also, you had earlier mentioned that in every cycle, there is some or the other learning. So, what is the learning from this cycle and how are we implementing the change?

Nikhil, thank you. We had almost INR 80 Crore of defaults in September because of the rains because we had rains in Bihar, Madhya Pradesh, Maharashtra, West Bengal, everywhere. But that default has already come down to INR 20 Crore. The rains of such kind of default normally will not be a long time. This is hardly one week or two weeks and next week money will come.

Similarly, even if in October we may have some difficulty, I am anticipating, Orissa, Bihar will have some default in the next two days. But still, this will come back. We really not worry much about it. So, whatever collection efficiency will remain stable even with that. So, we don't worry much about it.

The second one is about when the growth will come back. We estimate we should get at least four months of good, what you call, business momentum in this financial year. Therefore, we have calibrated about 8% to 12% growth potential for FY25. From the learning perspective, the data that Nilesh explained clearly shows that the default rates are coming from the new customers whom we acquired from different MFIs. Because if you see unique customer, even the new customer is behaving better. So, we need to improve the underwriting.

MFIN guardrails, one way it's helping us also. In some states, we have made more stringent rules than the MFI guardrails, particularly in newer states, so that we will not make similar mistakes

Page 14 of 19 what we did before. So, it should help us to retain the customers, which will help us to do well.

Again, data clearly shows that. So, the learning is to retain our customers and have better underwriting for the new customers in the new geography.

The next question is from the line of Shreya Shivani from CLSA India. Please go ahead.

I just wanted two questions. First, I am not sure if you have already shared this, but this 15.3% that you have shared the CREDAG +4 borrowers, as of Aug-24, what would that number be, say, a year ago or any like year ago, year-and-a-half ago, that will be useful.

And second, I wanted to understand that you spoke about, you split your branches into stable, non-stable, etc. So, is 70% of your branches that data that you shared out, does that mean that those are stable, and the group meetings are continuing over there and the rest 30%, there is some disruption in the group meetings? I just wanted to understand details of that 70% data that you were speaking about earlier.

Yes. For CREDAG, plus 4, we mentioned about 12% to 13%. A year ago, it was around 8%, if I remember correctly, but not exactly. Maybe Nilesh will check it and let you know later. And in the case of branches, I mentioned the centers, not the branches. About 70% of the centers are performing well; maybe it's distributed among the branches, but at least 40% to 50% of branches have a lower PAR. We are actually enabling those branches to conduct business better going forward because, at this point, we are actively managing them, as the surrounding system is not very strong. We’ve tightened controls there as well. Perhaps in six months, they will be able to manage within the 2% threshold. I think we are actually starting to work on this and motivate them to acquire customers and disburse more with a strong underwriting process.

And these customer centers, which are under stress, obviously the bigger chunks would be coming from the state of Bihar for you, right?

Not necessarily. Bihar portfolio itself is only 5% of ours, Shreya. It is only 5%. It won't make it work. It is distributed.

And sir, just last question, I wanted to understand if I just check the CRIF Highmark's data on Bihar state, the different buckets, it seems like the majority of the deterioration has just happened in the past 2-3 months, right? This is still June. The trends are still not so worse off as it looks right now. At least from your numbers, assuming your numbers are closer to the industry in that state.

Yes, it is the last 3 months. Our numbers have also increased in the last 3 months.

Next question is from the line of Bhavik Dave from Nippon Mutual Fund. Please go ahead.

Sir, a couple of questions, one is on incremental growth. And when we see, how difficult is it to add new customers? We have seen new customers addition drop out and for us to come back to the growth level that we thought about in FY26 maybe, in the current environment, how difficult is it to add new customers considering 6% of our existing customers are already in that 4+lenders plus 2 Lakh plus ticket size, right? So, even the existing customers will attract. So, a bit on how we are thinking about growth?

See, the growth will come from two buckets, one is from renewals, and one is from new customer acquisition. So, a large number of customers will come for renewal in the second half actually.

Even if we think about 6% of customers we have to drop, so we have already, if you look at, we are already acquiring more than 50,000-60,000 new customers per month with the new guardrails, right? So, it is a bit of scaling that normally will happen in the second half. You will have an increase in customers also, an increase in renewals also. These two things should take care of our, what I call, moderate growth, what we estimated.

Another question was, in terms of customer behavior, right? And we have understood that customers have a very high understanding in terms of their credit scores. Is this cycle wherein the incomes have not got impacted, but obviously some overleveraging has taken place. Do you see customers coming back and repaying as and when they have their money in their account rather than just going away? Is that recovery or behavior visible in this cycle because COVID incomes got impacted, so that during demon, this cycle is very different. So, how is the consumer behavior or customer behavior in this cycle, if you could just talk about that?

I think we need to wait and see. As of now, in the last three to four months, our experience says that actually moving forward is much higher than the recovery here. We have to see probably the next three months and see how it changes. As of now, it's mostly moving forward, the lesser recovery we are seeing.

The next question is from the line of Hardik Shah from Goldman Sachs. Please go ahead.

Sir, I have two questions. One is in the last quarter, you alluded that your ability to access customers was better in 2Q and hence you are receiving partial payments. How is that trending for delinquent customers? Number one.

And number two, if I see the fresh flows across states from Q1 to Q2, that has gone up across states. Karnataka has gone, for example, from 0.3% to 0.9%, Tamil Nadu from 1% to 1.7%, Maharashtra from 0.5% to 2.6%. So, what is giving you confidence that things will stabilize in 3Q and improve in Q4, given that fresh flows are continuing to come across states?

The access to delinquent customers is still good, but unfortunately, the recovery is a little less compared to earlier. We thought in the June quarter, we should be able to recover, but it is not exactly what we had wished earlier. It is not on expected lines. We expected to recover more.

Unfortunately, this kind of customer has multiple lending, and probably what we say over- leveraged or so, I think there are actually so many third-party interventions in many states. They are not repaying. So, less repayment is coming.

The second, what you said is, yes, there is an increase in the, what you call, Karnataka, Maharashtra also, but it is not as high as compared to the whole industry or compared to the other states. We strongly believe that this will still be controllable and then we will be able to manage much better. And already we keep checking the field, we keep checking the field information how it is moving. We are seeing the stabilization in both Karnataka, in Maharashtra and Madhya Pradesh and Chhattisgarh and even Tamil Nadu.

For example, today it shows a PAR 0 of 6.2%, but it has largely stabilized, with accrual additions in August, September, and October not exceeding those of the previous month. Therefore, we are observing similar trends in many states. We strongly believe that we should be able to turn things around in Q3 and achieve growth in Q4, unless something significantly different happens, as of now.

The ban made by RBI for a couple of NBFCs or also microfinance players, will that accentuate problems in some of the states that you operate in, in terms of liquidity and hence the problem that was accentuated by MFIN guardrails could kind of get accentuated even further?

It should not be because our overlap with any of these players is very, very low. Therefore, we don't see any issues because our maximum overlap is only with IndusInd Bank. We don’t have even double-digit overlaps with other players. So, we don't have any worries about that.

And sir, last question on your retail book, which you have grown a lot in this quarter, although it remains a very low proportion. Any color on the asset quality there? How is that trending?

No, that asset quality is very very good actually. I think, overall, within 0.6% of PAR 30 kind of thing. Sorry, it's 0.3%.

The next question is from the line of Gaurav Sharma from HSBC. Please go ahead.

So, two questions. One is, sir, we have seen that you have increased your branches, but in the number of employees, we have seen that they have decreased from the last quarter. So, what explains this divergence? That is number one.

And second, sir, you have mentioned in the medium term that the share of retail finance would be around double digits. And given your target of INR 50,000 Crore, so do you think you will be growing these segments more aggressively and their share may increase in medium term or you will stick to your previous guidance over the share of this retail finance portfolio?

Page 17 of 19 I will answer the second one. Now retail finance will grow as per our earlier model of what we said that by FY28, we would reach up to 15%. So, that is what the way we are doing. So, it will not grow too fast, but it will not grow too slowly also. 15% reaching by FY28 means we should reach INR 7,500 Crore, which is almost, from INR 1,000 Crore to INR 7,500 Crore is not small.

It will grow in speed, but the percentage will remain about 15%.

On the employees, it is a small aberration. I think it is about a few hundred employees who have changed. There may be more in trainees, which does not come in the number. So, there is no, what you call, worry about any employee attrition issues with that.

And sir, the hiring for these additional branches has already been done or it will come in the, like, the Opex.

Yes, normally we start with a few employees because once you open the branch, it takes at least two months' time to start the business. So, therefore, in that process, we will take employees and we will hire maybe one or two employees we would have hired first, one branch manager and one employee. Then within two months, we will hire more employees for those branches. Those are my only questions.

The next question is from the line of Nidhesh Jain from Investec. Please go ahead.

So, one question. So, the key issue at this time is over-leverage at the customer level. And this will keep on repeating with two, three years. So, what are we doing to make customers exclusive for us? Because the best asset quality experience is provided by the customers who are exclusive to us. So, from here onward, what are the steps we are taking so that more and more customers are exclusive to us?

See, our operating model is to always acquire and retain, Nidhesh. It is only in that process. It is only reevaluated that that is the best way of running this microfinance business. So, I think it is more of an underwriting improvement, which we already know, some states, they are actually stringent than the MFIN guardrails. Some more analytical abilities we have used. So, I think we will keep doing what we did and press additional underwriting process and high touch will remain as our core business model, which will help us to grow. That's why it is temporary, but yes, we will definitely learn some of these lessons and build a more robust process to retain our employees, our customers.

Sir, just to add that MFIN guardrails are also pretty aggressive. They are allowing up to four lenders. So, it is reasonably aggressive in my mind and up to INR 2 Lakh they are allowing. So, ideally we should be operating at much lower numbers, at least in the number of lenders internally.

Page 18 of 19 I think it should be calibrated. I think it's not that time to comment on that from my side, but we should look at the vintage customers of non-vintage and then we should work on a different model. As you see from some data, a vintage customer with high leverage, still it is good that there is no delinquency or less delinquency, right? So, we need to be a little careful in designing it, but we will at least this time it is important to start something. It may be 3 lenders or 4 lenders.

There was enough debate. Finally, we said, let’s start with 4 lenders. Maybe over a period we will move to 3 lenders also, depending on looking at the implementation and the results actually.

But the MFIN said that after Q3, they will again review and see the impact of the guardrails, whether to make it more stringent or to add more parameters, they would look at it.

The next question is from the line of Shreepal Doshi from Equirus Securities. Please go ahead.

Sir, my question was on how are the center meeting attendants trending in the month of October?

And from the qualitative aspect side, is there a, like, how is the inherent will to repay the loan, that aspect being discussed with customers from your RMs or your senior RMs who are right now going at the ground level probably?

So, centre attendance, there is no variation in October or September. I think largely we are more than 60%-65% in the southern market. So, it is a little less in the northern market. Now, it may be 50%-55%. There is always a huge difference. And willingness to pay, even if you take an average 50% or 50%-55%, your one-time collection is 95%. That itself shows there is a willingness to pay, correct? So, even they ensure that they are paying to the center actually.

So, if you ask me what is my door-to-door collection, it's only about maybe 3% to 4% other than everything that is paid in the center and on time. Therefore, sometimes we need to be giving that extra space to customers. Even though they are not coming to the center meeting, they are willing to pay and pay on time in the same center. They may have their own difficulties attending the center meeting. So, therefore, we are fine with that actually.

Just one last question. From these 27% centers which are relatively having higher delinquency, so which state would these be coming from or would be part of?

This is spread across all states. Higher the delinquency state would have a higher percentage. A lower delinquency state would have a lower percentage. Correct? But there were largely it is little higher delinquency states, like in Tamil Nadu, Bihar, Jharkhand, Rajasthan, and Kerala.

These are a little higher in terms of the percentage of centers.

And what strategy we have deployed to improve our collections there?

So, those centers need to be basically assigned to the better-experienced employees there. A higher proportion of centers will be handled by a more experienced employee to handle and motivate the customers.

Page 19 of 19 Thank you very much. Ladies and gentlemen, we will take that as a last question. I will now hand the conference over to Mr. Renish Bhuva for closing comments.

Thank you, CreditAccess team for the detailed answers. Thank you, everyone. We can now log off. Thank you very much.

Thank you, everybody, and Happy Diwali to everybody.

Thank you very much. On behalf of ICICI Securities Limited, that concludes this conference.

Thank you for joining us, and you may now disconnect your lines. Thank you.