Fewer Cards, Not Riskier Borrowers: TransUnion CIBIL CEO Bhavesh Jain On India’s Credit Card Slowdown
In this week’s The Core Report: Weekend Edition, Jain explains that credit card delinquencies appear higher due to fewer new cards being issued. Consumer awareness has improved, with many actively comparing credit card and personal loan options —indicating a maturing credit culture.
NOTE: This transcript contains the host's monologue and includes interview transcripts by a machine. Human eyes have gone through the script but there might still be errors in some of the text, so please refer to the audio in case you need to clarify any part. If you want to get in touch regarding any feedback, you can drop us a message on [email protected]
Hi and welcome to the core reports weekend edition, my guest for today is Bhavesh Jain, CEO of TransUnion CIBIL, the credit information bureau or those who keep credit or manage and follow your credit ratings or your personal credit ratings as you go through your life. Bhavesh, thank you so much for joining me. So, I think when it comes to credit ratings or of individual credit ratings and I do not want to confuse that with business or corporate credit ratings. What's yours?
Thanks Govind for having me on the show and my score is 805, which I think is good and the reason I am saying is that today if I go with our 800 CIBIL score, most of the banks would welcome me with a red carpet and give me the best terms and conditions. So, I think I have performed diligently over the last few years and made sure that I have a good CIBIL score.
And how did you reach that score? I know you did aim for a number and then work backwards but if I were to work backwards today, given I am sure a score which is lower than yours, what would I have to do?
So, like any regular Indian, took a car loan, a home loan, have a few credit cards and have made sure that I pay all my credit obligations on a monthly basis and make sure that I pay the expected amount, whether in case of a home loan, I pay the full EMI monthly, whether it's a car loan, I pay the EMI, whether it's a credit card, I pay the full outstanding amount. So, I am a transactor. So, from a bank's perspective, yes, they may not earn from my credit card because I swipe it, I use it and I pay it.
Similarly, on my home loan and my car loan, I pay it on a timely basis and I am very particular that I suddenly don't rush into taking too many loans. I space out my credit requirements and at the same time, I do a detailed planning to make sure that I don't get into a debt trap and I am able to live within my means. So, if you are taking credit, make sure you are able to pay it on a regular basis because that's the key and essence to getting a good credit history and a good CIBIL score.
And how long? I mean, for instance, let's say five years ago, things were not so good for me or I didn't have a proper job or my job was not consistent or I was jumping jobs and today things are much better. So, what is the timeframe that you usually use to assess backwards?
Okay. So, let me go a bit into the details. The CIBIL score ranges from 300 to 900. Higher the score, better is the risk profile of the consumer. Now, the scoring happens on three years of data, which includes aspects around credit offtake, credit behaviour, credit performance, types of credit being taken, the number of credit applications being made. So, these are the broad parameters on basis an individual is given a credit score or a CIBIL score.
Now, if you would have defaulted in the past, it's recommended if you start making your payments on a regular basis, you don't over leverage yourself, you stay within your means and you are able to pay not just your loans, but even your credit card outstanding. When you start making these payments on a regular basis, your score improves. And what we have done is that like every individual can come on the CIBIL.com, take their one free annual credit report and we have actually also added a score simulator to your point to make sure that the consumer knows that, okay, if I make this payment, if I foreclose this loan, this is how it's going to help me to improve the credit score. So, yes, it is possible to improve your credit score, but credit discipline is the key to improving your credit score.
And I'm going to come back to the credit discipline part, but are you a public service or not a public service?
So, we are a privately held company owned by TransUnion, majority shareholding. The regulations applicable for us is the Credit Information Companies Act, 2005. We are regulated by RBI and we have been in India, this is our 25th year of existence and I think so the common man on the street relates to our brand as CIBIL Score.
So, TransUnion CIBIL is our organisation name, but the most common known brand for our organisation is actually CIBIL Score.
Right. And my point is, you said the first time I can access my score, it's free and after that I pay you. So, how does your business model then work and who all are your customers, apart from those who are accessing scores?
So, yes, the customers are two-fold. One is the banks and NBFCs, second is the end consumers. So, banks and NBFCs for their underwriting will the credit reports of the individuals and the individuals come on CIBIL.com, take their free annual credit report and score and if they want to take it on a much higher frequency than the annual basis, they have to pay a subscription fees for that. But the banks, NBFCs give us data on a monthly basis and we match merge it, give a consolidated credit report, scores, multiple analytical solutions to banks so that they can understand the consumer profile better, they can underwrite the consumer better and at the same time the consumer bases their credit history and credit score can get the best terms and conditions in the country today.
As you do that, what's the mix of revenue today between what you earn from banks and the institutional customers versus retail customers who pay subscription?
So, we largely focus on our business revenue side, it's more towards working with banks and NBFCs because we want to make sure that for the consumer, we keep it very, very simple that we offer a free annual credit report and at the same time if the consumer wants at a higher frequency, then there is a small subscription fee but we largely like to engage with the consumer through a free annual credit report and what we work commercially largely is with the banks and NBFCs.
So, would you say it's 80-20, the revenue split?
Yes, or probably higher because our objective is to make sure that the consumer gets the credit report for free and we work commercially with the banks and NBFCs.
Right, okay. So, I'm going to come back to the consumer in a moment but tell us about the volume or the magnitude of data points that you're dealing with today and what is going on behind the scenes so to speak and I'll also ask you the mandatory AI question in a bit.
The banks and NBFCs under the Sikra Act are expected to submit data to credit information company on a regular basis. Now, the frequency has been changed and moved to fortnightly.
There are four other credit bureaus, all licenced by RBI, governed by RBI. So, the banks and NBFCs submit data to credit information companies at a higher frequency now. It's fortnightly data.
It used to be monthly, now it's fortnightly and once all the banks and NBFCs give in an automated way their customers, credit obligation and the credit information, those are synthesised. Obviously, it's a huge quantum of data, mix and match and when an individual pulls their own credit report, they get the entire credit history and the score. Similarly, when a bank pulls a credit report of an individual when he or she applies for a loan, they get a consolidated view.
Now, the quantum of data is quite huge. So, if you look at in terms of the ever credited and reported to CIBIL, would be around say 60 crore individuals data reported and the interesting part is that out of these 60 crore individuals, 27 crore are credit active population in India. That's when we look at 10 years, 20 years, 25 years since our existence in the country and as a company, we look at it, it gives a lot of sense of pride because the credit penetration has gone up significantly in India.
Today, when you go on the street and ask saying that okay, can you go and get access to credit, the answer would be absolute yes.
You are saying 600 million are people who have accessed the Sibyl system but 270 have loans of some sort right now.
So, 600 million would be individuals who have would have ever taken loan from the banking ecosystem because of which they have been reported to CIBIL. From the 60 crore individuals, 27 crores are credit active means they have either repaying it or they have applied in last 12 months which is a good healthy population.
So, 270 million is I'm assuming the kind of the base level for the data flow that starts. Each of those individuals and their multiple loans perhaps if they are multiple. Now, how many banks send you this data or NBFC send you this data?
So, we work with actually 7000 plus financial institutions in the country which ranges from public sector banks, private sector banks, NBFCs, regional rural bank, cooperative banks and the bulk in terms of the quantum of credit institutions actually cooperative banks because they operate in semi-urban rural locations and they enable neighbourhood lending.
So, you are saying of 7000, how many would be the regional rural?
So, regional rural banks have 40 plus but the bulk is cooperative banks. So, that would run into thousands because right from a single branch cooperative bank to a development cooperative bank to an urban cooperative bank to all categories of cooperative banks across the length and breadth of the country.
Right and so therefore, scheduled commercial banks, I mean the big bank, state bank, private, public all of that would be less than 100, I'm assuming.
Yes, yes, yes. So, the larger institution, the number of institutions are less but the quantum of consumers data they report is significantly higher but when you look at cooperative banks, the number of credit institutions are huge but at a bank level, the number of customers they report to CIBILis lesser but then they work in deep geographies. As I mentioned, they work on neighbourhood lending and it's important to make sure that that data gets reported to CIBIL because that's where the true financial inclusion happens because the larger institutions traditionally post COVID would focus on a credit-tested borrower.
It's still the cooperative banks who bring in the priority sector customer along with the public sector banks on the semi-urban and rural locations and a lot of interesting changes have happened post COVID in the credit market.
You know some of these borrowers who are far away, the bank that is lending to them, let's say it's a cooperative bank and this applies to anyone, they look at the CIBIL score as an indicator but they are not bound to act on the basis of that exact score, is it?
Yes, so what happens is that credit score is one of the parameters into the credit assessment. They will look at employment, they look at income information, they look at credit history, they look at credit score. So, there are multiple parameters they would look into a borrower when he or she applies to make sure that you underwrite the borrower in a holistic way and make sure then you decide whether to give or not to give a loan and to what extent to give a loan and to what tenure you want to give a loan.
Okay, so let me ask you a little bit about the data upload and the data dump. How does that happen? I'm assuming it's happening all the time but what's the interface? How do people connect to you?
So, two modes. One is when banks submit data to us, right now the RBI circular which came in effect from January 2025, it's quite progressive which ensures that credit institutions give data to credit bureaus on a fortnightly basis. Now, why it's important to get data on a higher frequency?
Because the bank gets a fresher data to understand the credit repayment behaviour of the consumer better. Same time for the consumer, the consumer makes a payment, it should ideally reflect in the shortest possible time so that when he or she goes and applies, their latest credit information is available for the bank to give the best terms of conditions. So, now this is fortnightly.
The way forward, it's actually moving from fortnightly to weekly to daily to real-time. That's the future which we see as the CIC in the country and that's the way forward because what happens the moment you move towards daily data reporting, both for the bank and the consumer it is beneficial. Because when a bank looks at the consumer's profile, they get all the information real-time.
That helps them to take the right credit decision. From the consumer's point of view, if he or she has made a credit card payment, a home loan payment, it should ideally reflect on a real-time basis. Now, the day when we move towards daily or real-time data submission, it's going to be a game-changer beneficial both to the lender and the consumer.
And when you look at data accessed by banks, we have already way ahead as compared to all the developed countries in the world. Today, a bank or the consumer can pull their credit report at a time taken less than a human takes to blink an eye. So, it's actually one odd second which is one of the best globally.
So, what happens is that today, we talk about quick commerce delivering in 10 minutes. You log on to any of the bank apps and if you're an existing customer, you click and you get a loan in 10 seconds.
So, that's where the credit discipline comes in saying that take credit when you need it. So, to come back to how banks talk to your system. So, banks work on core banking solutions. So, do core banking solutions connect to your system and tell me a little bit about that.
From their core systems, the data gets extracted through various scripts run by the bank and it comes into a standard format. So, the regulator has defined these formats of submitting data from the credit institutions to all the four credit information companies. It comes in a standard format.
It comes in an automated way, comes and gets posted into the predefined secured SFTP folders. It gets picked up from there. It gets uploaded, ingested.
One would look at what's the name, date of birth, gender, one of the national identifiers, look at the existing subject in the database whereas the reported, do the mix and match, run the algorithms and combine it into a single credit information report as and when it gets accessed by another bank. Everything happening in milliseconds or seconds but at the same time everything happening in an automated way.
Right and is this the software that drives all of this on your end? Is this a product that is globally used or is it something that you've developed?
So, it's a combination. We like to take the best practises from the global world, bring it to India but at the same time something which we can build homegrown like the latest technology stack on the civil side we have made it in-house locally and what's interesting is that on the MSME side or the small businesses side when we look at the credit infrastructure or the credit information companies ability to provide that information to banks is globally one of the best because India globally is the hub for small businesses and small businesses need credit the most in the country.
No, so just to go back to that software, so you're saying that the software that you use is a combination of product and something that you've developed internally and the reason I'm asking this is, so one of the classic complaints against CIBIL and maybe all personal information bureaus is that my information has not been updated. You know I've been paying on time and I paid and yet my bank is saying that it's not updated. So, why does that happen?
So, what happens is that banks get 30 days to report the data to bureau and that's why the regulator brought in the progressive circular saying that move from monthly to fortnightly. Now, if the consumer is paying on an X date and if the consumer have to wait for 30 days for their bank to submit data that can become a challenge for the consumer to get it. Yes, there are online tools right now available for banks to submit one consumer's information as and when they make the consumer payment.
So, we call it as like online maintenance where the bank bases the latest credit information can submit it at a consumer level. But the ideal case is to move towards daily or real-time data submission.
The time that is there…
So fortnightly if you ask me, fortnightly is better than the monthly earlier but the ideal would be daily data submission because you look at it the consumer makes a payment into a bank branch, comes out and in few hours the payment is updated.
So, let's say I had a big loan which I closed last week.
Yes.
So, will it be exactly 15 days later that my status will get reflected or is it a 15-day cycle that you follow?
No. So, a bank can send it on the same day when the payment is done and we will upload it. But banks generally send it with a frequency of 15 days.
But we have built in the technology where a bank can send it as many times they want to send it and we'll upload it.
And again I'm assuming the software that is looking at or the algos that are looking at this and saying okay here is so-and-so they paid their, repaid their big home loan on 20th of April and there's only now two loans remaining which are basically credit card which can get you know closed off anytime or closed constantly or consistently and the algorithm which sort of addresses that and gives a fresh score is also automatic operating real-time.
Yes. Okay. It's real-time because we want to make sure that the consumer gets everything basis what's reported whether in terms of the loan obligation, whether in terms of the repayment, whether in terms of the outstanding balances because the score gets computed real-time basis all these parameters.
So, if you foreclose a loan, if you make a advance payment or if you have made a delayed payment everything would get reflected real-time basis the submissions.
And I'm going to come to how the credit market in India is looking and your own study but before that you know you clearly have one of the larger banking financial system databases which is at 600 million at peak though 270 million are active.
Yes.
And in terms of computing power that you're using at any point of time which I'm sure is in the data centres how would you rate your requirements versus let's say a bank or any other financial institution or any benchmark that you use here.
So, I would say with pride and humility that today more than two decades when a loan gets applied for the banks pull the CIBIL report and score and we deliver that on a near real time as I said 1.2 odd seconds that's I would say globally one of the best and at the same time keeping the data in a secured manner keeping a data of a scale of this quantum and it's not data just on the consumer but it's a data on the MSMEs it's data on the microfinance borrowers as well.
So, all varied categories of data and managing this data with no unique identifier in the country. So, the unique identifier does not get reported to us it's the choice of the consumer and the lender that what identifier information they report to any of the four CICs. They can give the mobile number Aadhaar is not allowed but mobile number PAN number any of the other national identifiers and basis the submission we match and merge two subjects and deliver it real time.
And you mentioned MSMEs and you also mentioned microfinance. So, when we say 270 million active we are talking about individuals plus MSMEs plus or is it?
So, MSMEs are additionally so we have 3.6 crore MSMEs reported to CIBIL ever and out of those 3.6 crore MSMEs reported 92 lakh are credit active. Now that's where we see a huge opportunity for the MSME sector because out of the 7 odd crore MSMEs obviously there are different numbers quoted in NSS and other public domain information from 6.5 to 7 odd crore MSMEs, out of those 7 odd crore MSMEs 3.6 crore are ever created and reported to Sibyl and from those 3.6 crore 92 lakhs are credit active. So, when you look at it in terms of MSME only 12 to 15% of the MSME in the country are credit active and actually MSMEs need the maximum credit.
They contribute to 30% of the GDP but only 12 to 15% are credit active.
So, why is that? So, why is the credit activeness so low?
So, it's actually multiple things the credit penetration is low and then the credit activity is low because you look at from the lens of the MSME that the MSME borrower feels more comfortable to get an individual loan vis-a-vis on the MSME because there are higher turnaround times, there are additional documentation required when you take an MSME vis-a-vis when you take an individual capacity. In fact, we recently did a report called Builders to Borrowers along with NITI Aayog and there one of the data points which you highlighted is that women borrowers or consumers in India prefer gold loan and personal loan and then the third preferred product is business loan.
So, you look at it in terms of the preference of the women consumers in India actually gold loan, personal loan and then business loan because when you think from the lens of the women consumer it's actually the convenience because you can get gold loan in a much lesser timeline and at a significantly lower documentation and paperwork which obviously becomes very critical for a borrower when they are seeking loan.
Right and therefore we are also seeing Reserve Bank tightening gold loans separately. So, however when an individual borrows who is also attached to a small enterprise and a company in this case your credit roll, your CIBIL score hits both as in if I was an individual went to a bank or if my company went to the bank it would still reflect on my score right I mean it's not like a Tata Motors where the CEO's CIBIL score will affect Tata Motors rating but in a small enterprise in India both are interconnected, isn't it?
When you have the MSME and on the MSME when you have the individual associated with that MSME it gets reported under the relationship segment but in terms of your credit obligations are slightly different because they are two separate legal entities. Individual here the MSME is the proprietorship firm, partnership firm, LLP or private limited company.
The way the bank sees it if your individual score is low even if your company score is high the chances of getting that loan are equally impacted.
Yes, so when a bank look at giving a loan they look at the in case of small businesses they would look at the credit performance of the entity, the individual, stock, collateral, they would look at various aspects before they decide whether to give or not to give and at the same time use all those data points to even decide the credit limits and the tenure of the loans. So they look at every aspect of the individual entity associated before they take the credit decision.
Right, and what's been changing in this approach or in the way data has been analysed and to your benefit as well as to your let's say the your customers benefit and I mean the retail customer. So I last few years I mean the way you're using data the way you're extracting it and so on.
So I would say twofold one is that I'll first talk about the lens of the consumer. The consumer has become far more credit aware which is a very good thing because we have looked at is that a consumer who's credit aware their credit performance is relatively better as compared to a consumer who does not monitor their credit history. So we have close to 150 million consumers who monitor their credit history with CIBIL.
Now it's a very good number which means is that this segment who is credit aware it's very similar to health. If you know your parameters means you are conscious of how do you make sure that you have a good health. So why is this consumer doing it because the consumer knows that if I have a good credit history and if I have a good credit score I will get the best terms and conditions within the banking ecosystem in India because the rate of interest are linked to how the borrower performs on the credit history.
It also means that you know these are customers who have indulged in frenzy of borrowing in the post-COVID years and are now sort of trying to maybe either self-correct or trying to find out where they stand.
So actually the good thing is that a bank or an NBFC looks at credit repayment history, credit score, income, debt burden ratio they look at multiple aspects before they decide. Credit history, credit score is one variable into the credit underwriting and not the only variable. So it becomes a good parameter for the borrower to get motivation saying that yes I need to continue to pay on time.
At the same time the bank gets a idea about the individual saying that what has been the performance for the last few years. So it becomes a win-win.
So in these 150 million would you say the majority have good scores or the majority do not have good scores?
So actually the score gets computed real time. So if a borrower might have an X score one year back but now has paid or not paid the score might change. But I would say is that the consumer have become far more aware and at the same time we are working a lot with the consumers and the banking fraternity to make sure that we educate the consumer about the importance of having a good repayment history and a good credit score.
Because it is important the reflection of your credit repayment history needs to be known as a consumer and so that you can go and avail the best financial terms and conditions from the banks.
So let us come to the present for a moment. So you have talked about the architecture and you recently released a report where you have talked about how originations are going down because I am guessing there was so much frenzy in borrowing aided by let us say easy apps and so on and so forth. So what are the current trends that you are seeing in terms of retail lending borrowing?
So the credit demand and supply both have moderated and while what gets spoken publicly is the credit demand and supply moderation on the unsecured side which is your credit cards, personal loans but actually the moderation have even happened on the secured side whether it is home loan, whether it is car loans and our hypothesis is that because the when you look at some of the public domain data is that the home sales or the car sales have not grown by the same pace as expected and hence obviously if people are not buying more homes, more cars obviously the underlying credit demand for those purchases goes down as well. So yes credit demand and supply both have moderated but while what gets spoken about is personal loan but the biggest impact have happened on credit cards where the credit card supply or issuances have gone down, the delinquencies are slightly inched up. So it is one retail product which have got impacted otherwise when we look at all retail loan products including personal loans that in fact personal loan delinquencies are quite stable.
When we look at year on year after multiple changes in the marketplace the personal loan delinquencies have stabilised and secured loan delinquencies like home loan, auto loan have actually come down. It is only credit card which has slightly inched up but on totality retail loan portfolio qualities have improved year on year.
So whats the demographic inside this?
So interesting, very good question Govind. I would say is that while we talk about trade performance a massive shift has happened in the country post COVID.
So consumption loans which is your credit card, consumer durable and personal loans, the average age has come down which is very close to 30 years, 31 years to be precise closer to the median age of India and mortgages the average age has gone up, it is actually 41 years.
People are postponing their...
Yes and why it is happening is interesting data points. The consumption loan is largely taken by Gen Z's and the new to credit or the credit inclusion in the country is happening from two retail products. One is your priority sector agri lending and second is your consumer durable which is your phone loan.
In the past it used to be two wheeler loans along with agri. Two wheeler loans continue to bring in new to credit customers but the big quantum comes in through consumer durable which is a phone loan and why if you look at it...
More than 65-70% of phones are now bought on loans.
Yes because phone has become a big productivity tool.
It costs as much as a two wheeler.
Exactly, it is a big productivity tool. It is a good productivity tool. Individual comes out of post graduation or graduation gets into the job market.
The first thing they would need is actually a phone and you can get that phone on loan and that brings you into the credit fold. You have a good repayment and that is how you build your credit. In case of mortgages obviously post covid the underlying asset or the property prices have gone up and hence the consumer probably for the down payment is delaying the decisions.
So you look at it how it is panning out that you have the Gen Z's who prefer a shorter term loan because they do not want to commit for 10 or 20 years. It is the millennials or the people upwards of 40 years who prefer the long term loan or the mortgage loans and the interesting thing is that it is not the Gen Z's who are the fastest growing segment but the credit growth is happening higher in the semi-urban and rural location.
And why is that?
Because that used to be the credit under penetrated market and that is where the opportunity existed to bring in those consumers into the formal credit market.
But when you say semi-urban rural you are still talking about mortgages or mortgages plus consumer loans?
Consumer loans as well. So consumer loans in fact the highest growth has happened in the semi-urban rural locations and people less than 35 years age.
And I am assuming a lot of that has been driven by NBFCs.
So banks and NBFCs both. So whether it is the consumer durable loan, whether it is a two-wheeler loan, whether it is the agri loan. So these three actually bring in the bulk of new to credit borrowers.
And now that you mentioned that are you seeing differences in data that flows from when it comes to credit overview of the first to credit customers or even the older ones between the way NBFCs and banks are operating?
So their target segments are very different. Banks have the branch network focus. NBFCs will have a branch network and then feet on street model.
So there are different models evolving and everybody is competing for the best customer. It is actually the customer who has the choice to go to a bank or an NBFC or to a cooperative bank. So today the good thing is that customer has multiple choices whether he or she is in the rural location or urban location.
Because when you look at the number of branches being put up in semi-urban and rural location, it is a good number.
And between let us say within banks, between let us say state-owned banks versus private banks because we have clearly seen a lot of lending aggression by private banks. So is that showing up in your system?
So as I said the retail portfolio is actually looking pretty good on the secured side. Unsecured side, personal loans is stable. It is only the credit card.
And similarly, on the MSME side, the portfolio quality.
Yeah, I will come to MSME. How do you distinguish between personal loan and credit card for the purpose of computation?
So obviously when an individual goes and seeks a loan, the category is known that you are applying for a credit card, personal loan, home loan, car loan. So you disclose the product you need. The bank reports that data that this individual has taken home loan, car loan, tenure in terms of other details, the sanction amount, the disburse amount, the monthly payment, everything.
Got it. So if I come back to credit card, so you are saying there is a problem in credit cards.
Yeah.
What does that tell us about behaviour? I mean I know this is nothing new and various countries at various stages of development have gone through it. But are we in that phase right now where more and more people are taking credit cards or have taken credit cards and are sort of in that early phase of splurging and then maybe discovering that this is where they need to moderate.
So Govind, actually it is not more of the portfolio deterioration but it is more to do with the denominator effect that the number of new cards issued have actually reduced. And the way NPAs are calculated is your number of accounts delinquent than the total number of cards in force. So when you look at those, obviously the number of cards issued recently has come down.
Obviously that has been one parameter. Having said that, the credit card delinquencies have only inched up 30 basis point which is not as significant. The credit card outstanding are not that big in India still and also there is a little bit of awareness in the consumer that if I use a credit card, I might pay a X rate of interest.
If I take a personal loan, I might pay a very different rate of interest. So it is actually the consumer who is deciding that if I need credit card, do I use credit card or do I go and take a personal loan. So actually the consumer is deciding which is the preferred credit product.
And you are saying somewhere in this 150 billion active or hyperactive aware customers, people know even if they cannot do anything about it, they know what is going on in their credit lines.
Yes, which is a very good healthy indicator because we have seen that the borrowers who monitor their credit history do positive actions.
So I am going to come to what I call the mandatory AI question. But let us look at the data points that you are looking at. So you have 600 million accounts, 200 million of which are active.
Within that 150 are, you are saying, monitoring, directly accessing this thing and you have got 30 million small and medium enterprises. So what is the total number of data points that all of this is generating on a given day or a given month?
So the bank has the choice with a frequency of 15 days to report the data. So you look at it saying that it is the number of records actually run into billions.
Because that is my question because if suppose let us say I have in the last three years, if that is your active tracking period, I have taken one fresh loan but I am constantly repaying an old loan plus I may be spending on my credit cards. All these are data points for you.
Yes, yes. So whether it is a credit activity, whether it is 27 crores tomorrow, whether it is 30 crore plus or even if it is a historical loan, if there is an updation, the bank would submit that data. Okay.
So you look at the universe, say 60 crore would be the outer universe for the retail side to get reported at this stage as we talk. 3.6 crore of the MSMEs data to be reported. And the good thing is that as we bring in new to credit customer, this quantum will increase and as we increase or I would say reduce the frequency, the quantum will multiply.
Because you look at it, if we were uploading say x million of records, now with the frequency moving from monthly to fortnightly, you are doubling the quantum of records. Now we are recommending saying that the ideal state is that moving from fortnightly to weekly then to daily. So look at it, the entire credit population data getting reported.
So it is a huge quantum of data to be worked upon. But that is required because it is beneficial both to the credit institution and to the consumer.
And where are we in terms of the computing capacity to do this on your end and on the bank's end?
So I think so technology and India go hand in hand. So that is not an issue. Because the best technology brains in the world set out of this country.
So I am talking about compute capacity.
So that is the brain who brings in the, not just the compute power. See it is not just the raw compute power which matters. But how do you define your data architecture in an optimum way that you do not waste your compute power.
So your compute power and the way you process your data, the two go hand in hand and that is the most efficient way to run it. Because the raw compute power itself will not give you the best response times.
So what you are saying is that even if let us say your data dump doubles or triples in size because it goes from fortnightly to weekly to daily, then your system is ready for it. Now it is only really up to the bank and the 7000 institutional customers.
And the good thing Govind is that even the banks are interested to move towards weekly and daily data submission. Because everybody in the process benefits. The bank gets to let the world know.
So what stops them today?
It is just a matter of time I would say.
So it could happen that let us say half your 7000 banking or financial customers start uploading every day, the others take longer or do all have to operate on a 15 day basis.
So if the regulatory requirements change, like the regulatory requirements now mandates fortnightly. If there comes a regulatory mandate to move towards daily data submission, then all the 7000 institutions will do it.
But fortnightly I am assuming is minimum.
Yes, you can do it at a higher frequency also.
Are people doing that?
Yes, some of them do.
Okay, so how often are they doing that?
Some of them are actually giving their new accounts open on a daily basis.
Okay, so this is already happening?
Yes. So in a smaller quantum, but I think so it is required that we move towards daily data submission.
Okay, that is interesting. So to the AI question now, so you described the data points that you are dealing with on a daily basis or a fortnightly basis. So what are the kind of AI applications that you are seeing today to improve the quality of your own insights and insights for your customers?
So it is multifold. When banks access data, they access on their existing customers. When a customer applies, we use various machine learning models, AI models to assess the borrower.
Obviously, the CIBIL score being the flagship score which the end consumer understands. And there we try to use the latest modelling techniques, but at the same time keep it very simplified because the end consumer needs to know that it is the repayment history, it is the loan obligations, it is the number of applications you make. So you put in all those parameters.
So one has to make sure that you have the balance where the end consumers understand. At the same time, you help the lenders to assess in a very sharp manner because what is the objective of all these credit models is to help the banks understand the risk better or help them manage the risk better of the customers coming through the door as well as the existing portfolio.
But there is an algorithm, there is a base algorithm. It is almost like the Koch formula which is yours where you eventually add up to that 800. Now my question really is what is changing in that?
Suppose let us say thanks to AI, you are able to attribute some higher score for behaviour which you could not do earlier or were not doing earlier or are you saying that the formula like the Koch formula is the same all this time, all these years that you have been around?
So actually, I would say it is the modelling technique and the modelling technique is a function of the modelling variables and parameters and it is also a function of the underlying data. Whenever there is a macroeconomic change, you need to rebuild your score. So we have already had 3 versions of scores in last 20 odd years in India and it is important that every 4 to 5 years, the credit information companies need to come up with a new version of the score because the underlying macroeconomic or the customer base has shifted.
Because if there is a population shift, you need to make sure that you incorporate the population shift. At the same time, it is a very simple thing is that you want to look at a credit repayment history that will continue to be one of the biggest parameter every credit model built in the country or globally because the underlying thing is that you want to assess the credit worthiness of the consumer.
But what has changed? So let us say when you say this model has changed in 3 times in the last 20 years, between the last time and today, what are the changes even within the whole architecture?
So I would say the number of data points available. If 20 years back, the data points available were X, now they have grown multiplied. One, second is the frequency of the data available.
Third is the depth of the information available that you have today not just the large bank submitting data but even the corporate bank submitting data. So what you have the width, the depth and the availability of data and the frequency of data that has really really changed many fold. The quantum of data available to build a model and to run a score on a real-time basis, everything has changed in last two decades.
But when you say macro shift, what is the macro shift that you are referring to?
So if a consumer's credit in 20 years back is to be concentrated around say metro and semi-urban, now the geography spread is very different than two decades back, one. Second is that 20 years back, an individual took largely say only a home loan and to some extent a car loan and urban customers going for credit card only. Now the biggest credit growth is happening in the semi-urban rural locations where the type of loans being taken are gold loan, home loan, two-wheeler loan, consumer durable loan.
So the different types of loans been taken have changed completely in last two decades.
But how does that influence what my score could be or my anticipation of what my score could be?
So score is simple of what data comes in. It is like very similar to food that you can make a dish based upon the ingredients available. If you have multiple ingredients, you can make multiple dishes and you can have a multi-course cuisine.
Same way if you want to build a score or a model, if you get multiple data points, if you get sort of across geographies, across age brackets, if you get across product types, it is very similar to food dish being made.
Right and that will taste different in every household as they say in India. So to that extent, I mean can I shop around for other credit ratings, individual credit ratings because as you said there are three others.
So the law says that every individual can go on to any of the CICs and get one free annual credit report free. So if an individual has taken say home loan, car loan and say a credit card, all three would reflect largely in all the four credit reports. So the credit reports will reflect the credit obligation across the four CICs.
An individual can decide to choose which credit bureau they want to go for.
Right, so let me come back to the latest figures that you have. So credit demand in your data says that it is slowed down and you are also projecting that it is slowing down. So what are you now seeing ahead?
I mean we are now in let us say the beginning of the new financial year, that is 25-26. How are things looking?
So for the January-February-March quarter, obviously the credit demand is not the traditional GFM credit demand and it is not just the unsecured credit demand and supply which is moderated. Even the secured credit demand is moderated. Obviously one knows that historically whenever there are rate cuts, you will have the rate of take going up because obviously it eases the available liquidity in the market.
There is a rate cut passed on to the consumer. The consumer finds it sort of interesting to go and apply for a loan. So the expected next 12 to 24 months is based upon the macro conditions, the pass on of the rate of interest to the consumer.
We do expect saying that from year on it to take forward and move upwards.
So what should people do? Basis your findings. What can banks do?
I mean not that you are telling them what to do but what could they do? And customers in any case you have talked about which is a more active, you know keep tracking your score and keep repaying and so on and make sure that you are debt free at least for three years. I am assuming before three years whatever happened is history.
So it will reflect in your credit history. But the credit score is which is the last three years of repayment history.
Okay, that is good to know. So I am going to use a question which someone sent us. So Indrani Dasgupta asked us, she read a piece that Reserve Bank is reconsidering the fact that even if one defaults non-loan EMI, CIBIL won't hit. So I don't know, does it make sense?
Non-loan EMI?
Yeah, so maybe it is an EMI which is not connected to, maybe if it is not connected to a bank, maybe or is there any, yeah.
So I will explain it. So if it is a credit institution, whether it is a bank, NBFCs, cooperative banks, RRBs, any of them, if you have taken a loan, okay, and you are paying it, whether it is a loan or a credit card, in an individual capacity or in a non-individual capacity, that would get reported to CIBIL or any of the four CICs on a fortnightly basis. So that is a mandate in terms of the reporting by the credit institutions to the four credit information companies.
Okay, and if you were to look ahead now, I mean from a more macro point of view and looking at, you know, how internationally things work. So for example, you said that our turnaround is amongst the fastest in the world. I would also look to like, you know, why is it that it is faster or why are others not as fast?
Is it because the technology systems are newer here or the load that you are dealing with is different? But also, what is the future for personal information sharing looking like?
So I would say as CIBIL, we have always been committed to bring in the best technology and yes, why it is important because there is a very strong linkage. Credit growth and GDP growth are linked and we want to make sure that for the credit growth to happen, it is not just that response time of one second, but how do you make sure that you are able to get, upload, ingest and deliver the quantum of the data which India has. For India to grow, credit growth is important and we want to make sure that the best technology and the most secure technology is available for the end consumers and banking.
And yes, going forward, we see the quantum of data multiplying significantly. As I said, the way forward is to move from fortnightly data submission to daily data submission. So the quantum of data is only going to increase and it is our sort of commitment to make sure that we continue to invest and bring in the best technology and yes, because as you can see that the payment ecosystem in the country and the credit ecosystem in the country and we should be all really proud as Indians, as India stands out one of the best globally.
You can make a payment at the click of the button and at the same time, a consumer if they need credit or if an MSME need a credit, they can get it very easily. They don't have to wait for weeks and months. It's simply basis available of data of the consumer that the credit institutions decide to give a loan.
Right. And last question. So in terms of policy, and you've talked about one critical input, which is the frequency of data, which should be speeded up and so on.
Is there a wish list that where you feel organisations like you can be more empowered, which in turn will hopefully help consumers be more empowered? Is there anything that you're looking out for?
Yes. So I would say Govind, it's actually the non credit data. Because when you look at India's ecosystem, you have a lot of consumers sitting in the semi urban and deep geographies, rural locations specifically, they may not have a great footprint.
How do we bring in those individuals and small shopkeepers, businessmen into or the agri related farmers into the formal credit sector? Because for the bank or NBFCs to take a credit decision, they need information about those individuals. Now, if somebody is borrowing for the first time, you don't have credit footprint.
And for us to build any reports or scores, we need non credit data or alternative data. So information like mobile bills, electricity bills, insurance, subsidy, all those data points can come really handy to actually score a new to credit borrower, and actually bring in the borrower or consumer in the remotest of location into the formal credit sector. Because the moment you give a credit information, whether it's not just the loan, but how a borrower pays or a consumer pays on mobile, on electricity, the frequency of insurance payment, so and so forth, banking information.
And one interesting data point when you look at it saying that today, a small shopkeeper may not have access to credit or may not have taken a loan. But thanks to UPI, that shopkeeper will definitely have a bank account and a UPI handle. Now, that bank statement information is good enough to provide a new to credit score, which will enable that small shopkeeper to get access or the panwala to get loan from a bank.
UPI is there, but those data is not available to create information companies to actually build credit scores and give it to banks and NBFCs to underwrite the consumers.
The banks have to share that with you?
Yes, the banks have to share.
And that's only triggered when someone applies for a loan, but not to actually...
Yes, so the current framework allows us to get only the credit data. As you asked, what is the wish list? The wish list is to get non credit data or alternate data, so that we can score the new to credit borrowers.
And as a hypothetical example, if let's say Vodafone or Jio wants to share data with you for whatever reason, there's nothing stopping that?
The current regulations don't allow them. The regulations which govern us, the regulations which probably govern the respective entities, we'll have to modify or go for changes in the regulatory.
This is happening in other parts of the world?
Yes, in some parts it does happen that the CICs might get access to a non-credit data as well.
But on demand, I mean because the consumer accepts or is it because they're sharing all the time?
So it's a function of the consumer and the lender both. If the lender has that data or any of the ecosystem players have that data, they can submit it. Obviously, the consumer should be fully aware and with the consent of the consumer that they give this non-credit data to the bank and then the bank submits it or any of the utility companies submit it to the CICs.
And if I've been paying my electricity bills on time and my mobile phone bills on time and within the day, does that make me a better customer for the purpose of your starting point?
So yes, whether it's a loan or it's a utility payment, the future prediction generally happens on the past behaviour. If one has paid for one year, two year, three year, five year, then you know in all likelihood this consumer has a high chance to pay in future as well.
Bhavesh, it's been a pleasure speaking with you. Thank you so much for having me here.
Great, thanks for having me. Thank you.
