00:00 So good morning everyone,
00:01 and good afternoon.
00:02 Uh,
00:03 thanks so much for joining this seminar today,
00:05 uh,
00:06 that the WDR 2024 team is hosting.
00:10 It's looking at,
00:11 uh,
00:11 India's uh digital payment systems,
00:14 and,
00:14 uh,
00:15 this is a really important seminar.
00:17 Because
00:19 too often the discussion of uh around
00:23 technology
00:24 and,
00:24 and digital access has just been around applications
00:28 and expansion of um of use,
00:31 and I think we're very
00:34 Happy that today we have Professor Pulag
00:36 Ghosh from the Indian Institute of Management.
00:39 Will provide us some very new research on looking at
00:44 the extent to which a person's digital capital
00:47 can be used to expand access to credit,
00:51 and uh this is an important
00:55 paper for many reasons.
00:57 One of them is,
00:58 of course,
00:59 uh,
01:00 you know,
01:00 uh analytic skepticism,
01:03 whether expansive expansion of digital services
01:07 really has a multiplier effect
01:09 in terms of using,
01:11 in terms of accessing,
01:12 uh,
01:12 financial capital.
01:14 Uh,
01:15 so,
01:15 uh,
01:16 Professor Ghosh.
01:17 He is professor of Decision Sciences,
01:20 uh,
01:20 uh,
01:21 and,
01:21 uh,
01:22 at the Center of Public Policy at the IAM Bangalore.
01:25 He's been a distinguished fellow at Nithi IU,
01:29 which is the Indian government's highest decision-making body,
01:33 uh,
01:34 or policymaking body,
01:35 and has been consulting with governments on various issues of technology,
01:41 including AI,
01:42 big data,
01:43 FinTech,
01:44 digital intelligence,
01:45 and such.
01:47 Uh,
01:47 and
01:48 today he'll tell us a little bit about his work on,
01:52 uh,
01:53 how open banking expands credit access.
01:56 Professor Ghosh will speak for about 35 minutes if that's OK,
02:01 Pilla,
02:02 and we can,
02:03 and then we'll turn to questions and answers.
02:06 But in doing so,
02:07 if you have questions that come up,
02:10 please add them to the chat,
02:12 and then I will try to kind of
02:15 Ask uh Professor Ghosh to reflect on those questions as he,
02:20 as he presents
02:21 the paper.
02:22 So,
02:23 Professor Ghosh,
02:24 over to you.
02:26 Thank you,
02:26 Shamik.
02:27 Uh,
02:27 let me just pull up my slides.
02:33 OK.
02:34 Thank you for,
02:35 uh,
02:35 uh,
02:36 Summit for arranging this.
02:37 Uh,
02:37 I'm very happy to be,
02:38 uh,
02:39 presenting this new research of mine,
02:41 uh,
02:41 which is actually a continuing effort.
02:43 I have done a few more along the line,
02:46 but this is,
02:47 uh,
02:47 a much more broader picture,
02:49 uh,
02:49 in,
02:50 in this phase.
02:51 Uh,
02:51 so these are,
02:52 uh,
02:52 jointly done with my co-author Shashat,
02:54 uh,
02:55 Nirupamma and Manjo.
02:57 So the basic idea is that you know if we look
03:00 at credit growth overall as well as the financial inclusion,
03:04 both are very,
03:05 very important for any policymakers in any country.
03:09 Now,
03:11 Various countries,
03:13 including United Nations Sustainable goal,
03:15 if you look at all the seven charters,
03:17 financial inclusion,
03:18 you know,
03:19 is definitely part of it.
03:21 Now,
03:23 so far by financial inclusion,
03:24 people thought that having a bank account
03:27 will give the true financial inclusion,
03:30 and when I say true financial inclusion,
03:32 it's more of a credit access.
03:35 But things haven't
03:37 gone
03:39 that much.
03:39 For example,
03:40 in India by July 2016,
03:43 99% of the Indian household
03:46 had a bank account,
03:48 and,
03:48 and,
03:49 but that did not really put the credit expansion that much.
03:53 Now in between,
03:54 there is a lot of promise the fintech players in the market came in with.
03:59 And and the major reason why just having a bank
04:02 account did not expand the credit access as expected is because
04:06 the majority of these people who got the bank accounts,
04:09 they don't have a credit history.
04:11 And in the conventional wisdom of marketplace like banks to give
04:15 credit is that you need to have a credit history.
04:18 Now the Fintech guys came in between and they shown a promise by saying,
04:22 OK,
04:22 if you don't have a credit history,
04:25 maybe I can develop a credit score based on alternate data
04:28 like mobile data or any other thing,
04:32 but that also did not scale up or did not
04:35 lift up to the expectation as papers from Craig Butcher,
04:38 Gopal,
04:38 Manisha,
04:39 and they have shown.
04:40 So now the policymakers are thinking that OK,
04:44 if everybody has a bank account,
04:46 does then a digital payment history or digital public infrastructure,
04:51 not public,
04:52 digital
04:53 financial infrastructure
04:55 helps in getting credit expansion.
04:58 Now why the
05:00 policymakers think that's that's a you know a necessary step to do
05:04 is because
05:05 if there is a digital payment history done,
05:08 then even for banks,
05:10 if not for fintech,
05:11 even for banks,
05:12 the cost of
05:14 you know,
05:15 screening the borrower
05:17 becomes really,
05:17 really low,
05:19 and having a digital infrastructure,
05:21 especially financial and payment history infrastructure.
05:25 Uh,
05:25 reduces the cost of onboarding customer.
05:28 For example,
05:28 in 2014,
05:30 uh,
05:30 on an average,
05:31 the cost of onboarding a customer in India was $12
05:35 and that came down to 60 cents as we,
05:37 as we speak,
05:39 right?
05:39 So then the question becomes that
05:41 does a payment rail,
05:43 which will have all the digital transactional history of a customer,
05:49 uh,
05:49 help?
05:49 Now this can be either in a private way,
05:52 for example,
05:53 Uh,
05:53 the credit cards do or the banks do,
05:55 or this can be in a public way which India has done
05:59 or Brazil has done.
06:00 So India has UPI similar
06:02 in spirit.
06:02 Brazil has done peaks around COVID time.
06:05 So
06:06 this is the base of our paper.
06:09 However,
06:09 there are very limited understanding
06:11 that whether having this payment trail
06:14 does help in credit expansion.
06:16 So in this.
06:18 Research,
06:19 but we are focusing
06:20 on exactly that.
06:23 Now what is open banking?
06:25 I'll,
06:26 I'll
06:27 be very quick in this,
06:28 I guess all of you kind of know about it.
06:30 At the heart of open banking is the customer.
06:33 What does it mean?
06:34 It basically means that
06:36 customer owns the data,
06:38 although it may be stored by the bank or the major financial institutions,
06:42 but customer at the end of the day owns the data,
06:45 and customer
06:46 can consent to sharing the data to somebody they feel comfortable
06:51 for something.
06:52 Now
06:53 there are about 87 countries as per the IMF report in 2021
06:59 who are trying to delve into the open banking,
07:01 and 47 countries already adopted some amount of open banking,
07:05 although the degree of intensity varies by countries.
07:08 For example,
07:09 in the UK,
07:10 7 largest banks are mandated to share the data.
07:13 It's not across all the financial institutions.
07:16 Singapore,
07:17 US,
07:17 there is no explicit mandate yet,
07:19 but it's more like industrial-led mediation.
07:21 On the contrary,
07:22 India adopted fully the open banking customer permission data sharing,
07:27 so any customer
07:29 can share the data
07:30 to anybody
07:32 from bank to fintech,
07:33 fintech to bank.
07:35 It's completely
07:36 complete open banking
07:38 now.
07:40 How does uh
07:42 So what is the research question we are after?
07:45 Now,
07:45 what we found out of all these countries who
07:47 are at various stages of adopting open banking,
07:50 India has a very unique open banking structure.
07:53 It has two very important pillars in their open banking structure.
07:56 One,
07:57 it created a digital public payment rail.
08:01 Uh,
08:01 now it's public in the sense government or in
08:04 the sense National Payment Corporation of India made it
08:07 and opened it up to
08:08 any third player.
08:10 So it's not controlled by government.
08:12 They just made the infrastructure and made it open to anybody.
08:15 Now this infrastructure
08:18 does something very interesting.
08:20 If I transact on that infrastructure,
08:23 then I will be generating a real-time digitally verifiable financial transaction
08:30 free of cost.
08:32 And it's stored
08:33 by various parties.
08:35 Second,
08:36 because this payment history is being stored free of cost,
08:39 both for the merchant as well for the retail customer,
08:42 now I can share this to any financial intermediaries.
08:45 The Majority of the open banking talk about the Part B of the scenario,
08:49 which is the customer can share the data.
08:52 India has gone ahead by saying
08:54 it's not only important that you just share your banking transaction data,
08:59 you can even generate in real time
09:01 without going through the banking.
09:04 Now how does this open,
09:05 so the question we are after is,
09:07 if this is the scenario which India has,
09:09 in this case India's UPI,
09:10 Unified Pay Interface.
09:13 Does it help in credit access,
09:14 expansion,
09:15 credit growth?
09:16 If it does,
09:17 then
09:18 whom it does for?
09:19 Does it for
09:20 new to credit borrowers,
09:22 which is more financial inclusion angle,
09:24 or it's more to the intensive margin,
09:26 which is ex ante included borrowers are getting more benefit?
09:30 Now which financial intermediaries are actually dominating this effort?
09:34 Is it Fintech
09:35 or it traditional banks,
09:36 or both are getting
09:38 positive spillover?
09:40 And does the distributional impact,
09:42 you know,
09:42 distribution of the borrowers changing,
09:44 right?
09:44 So these are the
09:46 broad set of questions we are after.
09:49 Now,
09:50 in any empirical research,
09:51 data is enormously important.
09:53 Uh,
09:54 so,
09:54 so I'm putting the,
09:55 uh,
09:56 we have collected massive amounts of data from 6 different sources to study this,
10:01 and these are all pan-India data.
10:03 So we got the very detailed credit
10:06 registry data on retail loans from TransUnion CIBil
10:09 from Q1 2016 to Q4 2019.
10:13 We have the data beyond that,
10:14 but because of COVID,
10:15 we got it at 2019.
10:17 Which is more of a liability side of the data,
10:19 but the data is not at the individual level,
10:21 it's aggregated at the PIN code,
10:23 which is like zip code in India,
10:25 and month,
10:26 but it is cut
10:28 across lender type,
10:30 across borrower type.
10:31 We also got
10:33 the monthly UPI or digital financial transaction aggregated at zip code from 2016
10:40 to 2019.
10:42 We got the deposit data in bank branch within a
10:46 zip code for all the bank branches of India.
10:50 We got the bank accounts data.
10:51 If you're aware,
10:52 there is there is something called Prime Minister Janathan Jona account
10:57 which is basically
10:59 started in 2014,
11:01 which tells the banks with an RBI mandate.
11:03 That if a person goes to a bank to open an account,
11:06 even if they don't have documents,
11:08 banks are supposed to open the account,
11:10 and the customer can provide the required documents in the next 6 months
11:14 post account open,
11:15 and more importantly,
11:17 they can have under GDY accounts
11:20 zero balance,
11:21 so not necessarily they have to have non-zero balances,
11:24 right?
11:25 So we've got that data.
11:26 We have the telecom towers data,
11:29 and when I go to the empirics,
11:30 I'll show why that is important.
11:32 And
11:33 later on we found we got data from a large fintech firm which actually
11:38 lends based on this UPI transactions data
11:41 just for more digital verifiability.
11:44 OK,
11:45 so what do we do,
11:46 combining all this data,
11:47 we have 5 key results.
11:50 First,
11:50 we show that
11:52 financial inclusion or credit inclusion.
11:55 Just by bank account is not enough.
11:58 This bank account
11:59 needs to be powered
12:01 by this kind of open payment infrastructure.
12:04 So our
12:05 main result is that a bank account
12:08 powered by
12:09 this digital payment history data
12:12 is a deadly combination for credit inclusion.
12:14 Second,
12:15 we see that there is a manifold increase in
12:17 credit to the new to credit and subprime borrowers.
12:20 FinTech leads the growth,
12:22 but not necessarily that bank and MCFC are far behind.
12:26 Banks are also expanding credit,
12:27 not necessarily to the marginal or underbank borrowers,
12:31 but even the prime borrower base in banks are increasing.
12:35 Regions which were previously unbanked,
12:38 but they came in through the JDY accounts,
12:41 they got the maximum increase in the credit led by the fintech,
12:44 and all these are happening more in regions where the internet
12:48 is fast
12:50 and very cheap.
12:52 So because if you remember this digital payment creation history
12:55 creation needs internet because your mobile needs to have data
12:59 capability
13:00 to transfer the funds,
13:02 OK,
13:02 so.
13:03 Now
13:04 there has been literature wise there has been a theoretically ambiguous result
13:09 so far one or two papers on this angle,
13:12 whereas one paper
13:14 shows that
13:16 cross-country
13:17 open banking analysis,
13:19 which shows that
13:21 countries which have been adopted open banking in a much more in-depth way,
13:25 the funding for the Fintech through VC has gone up,
13:28 but they did not find much for the financial inclusion.
13:32 Other papers,
13:34 some of them show that
13:35 it's at the end of the day,
13:37 it's open banking creates a competition between bank and fintech,
13:41 and it depends who can screen the borrower data much better,
13:45 but they did not find any empirical result to support it.
13:48 Another paper shows that in the developing market,
13:51 probably open banking should do financial inclusion,
13:55 but in the more developed market
13:58 this brings more
14:00 competition,
14:01 and banks may not have incentive
14:04 to actually do this because
14:06 what can happen is
14:07 if
14:08 This data
14:09 sharing through open banking starts getting more matured,
14:13 then banks will lose business,
14:15 their profitability,
14:16 so they don't see enough incentive for them to store the data
14:20 and develop much more
14:21 high-tech screening technology
14:23 for their screening the borrowers.
14:26 So our primary contribution are two.
14:29 First,
14:29 we introduce a new channel of credit inclusion which is saying that
14:33 we need digital payment history data
14:36 added in the bank accounts.
14:38 That's a combination which can lead to formal credit.
14:40 And second,
14:41 empirically,
14:42 we are the first study
14:44 to show that this open banking
14:46 through open public digital
14:48 payment infrastructure
14:50 actually
14:51 increased credit market
14:53 across institutions
14:55 across borrower.
14:56 OK,
14:57 now,
14:57 why India?
14:59 So India is actually a very unique setting.
15:03 We can safely say India is now a world leader in digital public infrastructure,
15:07 and it started way back.
15:09 A lot of things came in place starting from 2008,
15:12 including Adhar,
15:14 then Jandan Jojna,
15:15 then mobile,
15:17 then UPI.
15:18 Currently India is working on the account aggregator,
15:21 so a lot of things came in place which we call together as India Stack.
15:25 In all these we are now focusing on unified payment interface,
15:31 which I will come in a minute,
15:34 which basic premise of that is a free
15:37 interoperable
15:38 payment system,
15:39 free for this is a very important part.
15:41 It's free for both financial intermediaries
15:44 and the customers,
15:45 OK.
15:47 All right.
15:48 Now,
15:48 what is unified payment interface?
15:51 So this is like
15:52 a
15:53 public
15:54 payment trail
15:55 done by the National Payment Corporation of India.
15:59 And
15:59 it has certain features.
16:01 So for example,
16:03 let's let's look at how it works.
16:05 So this is a typical flow
16:07 of a UPI.
16:09 So
16:10 I can download a mobile app for UPI.
16:13 Now
16:14 that mobile app for UPI will have,
16:16 so UPI is the payment rail.
16:18 It's a public infrastructure.
16:20 Now any third party can develop a UPI handle.
16:23 Now who are these third parties?
16:25 This can be Google Pay.
16:26 This can be Amagen Pay,
16:27 this can be,
16:28 you know,
16:28 those are the third parties.
16:30 So those third parties,
16:31 let's say for example,
16:32 I download Google Pay,
16:33 now I need to connect
16:35 with the UPI,
16:36 which is NPCI's Payment rail.
16:37 I connect that.
16:38 And I want to send money either to a merchant or to let's say Summit.
16:44 So this is how it works.
16:46 I enter the amount
16:48 and the payee VPA,
16:49 Pay VPA can be the mobile number,
16:51 it can be the QR code,
16:53 then the request is sent to the UPI
16:55 or NPCI
16:57 databases.
16:59 They
17:00 first check if you are the authentic person.
17:02 Once that authentication is done,
17:05 then
17:05 the address and authorization verifiability happen for
17:09 the account I'm sending the money.
17:11 Once that happens,
17:13 then it contacts the beneficiary bank.
17:16 Credit the beneficiary bank,
17:17 remittance happen,
17:19 and come back to the payer by saying this much amount has been transferred.
17:23 This entire process of transferring money takes
17:26 somewhere between 20 seconds to one minute.
17:29 It's that fast.
17:30 OK.
17:31 Now how does it look like?
17:32 So let's say I have a UPI handle like
17:35 GP.
17:36 If I have GPay,
17:37 there are steps through which I
17:39 can download this.
17:40 First,
17:41 I need to connect to my Aadhar.
17:43 Once the Aadhar happens,
17:44 then I need to put my phone number.
17:46 Then a one-time password comes.
17:48 This Aadhar goes and checks in the other databases of the government of India,
17:51 authenticate it
17:53 through your phone number,
17:54 then it will.
17:55 Asks to connect a bank account.
17:57 Remember for UPI,
17:58 a bank account is a must,
18:00 and India has done that before UPI
18:02 through Jandha Jona.
18:04 Now once the bank account is done,
18:06 several process a password,
18:07 then I have
18:09 a UPI account
18:11 through G Pay handle.
18:13 Now
18:14 once that is done,
18:15 let's say Google Pay looks like this.
18:17 So what has happened
18:18 here?
18:19 You have a QR code.
18:21 If you want to go to a margin and pay something,
18:23 you can just scan the QR code.
18:25 You can pay
18:26 anybody to pay
18:27 all your contacts in your mobile phone,
18:29 and so on and so forth.
18:31 And
18:32 if you scroll down that page,
18:33 landing page,
18:34 you will see all these things.
18:35 For example,
18:37 if you look at the very right hand side,
18:39 you can check your civil score
18:41 instantly.
18:42 You can show the transaction history for as many days as you want immediately.
18:46 You can check your credit bank balance almost 24/7.
18:50 And then
18:51 you can also pay your bills.
18:53 You can recharge,
18:54 pretty much do everything,
18:55 OK,
18:56 directed from the bank account
18:58 now.
19:00 If you look at the middle graph,
19:02 I can add
19:04 as many payment methods as I want.
19:06 I can add as many bank accounts I want,
19:08 I can add my credit card,
19:10 I can add my debit card.
19:11 Everything can be there.
19:13 And once a transaction has happened,
19:15 then I'll get a screenshot like this last graph here
19:18 by saying
19:19 I have paid this merchant Mac Foods ₹10
19:23 and it has gone from my
19:26 ICSI bank account,
19:27 which has this
19:29 UPI transaction ID.
19:31 So that's in a nutshell
19:34 what UPI is.
19:36 Now this is the infrastructure part of it,
19:37 but why it is becoming so
19:40 fascinatingly attractive to customer and market.
19:44 First,
19:44 it's interoperable
19:46 in the sense
19:47 that I can have,
19:47 let's say 3 bank accounts,
19:49 and through this UPI handle I can transfer money between my bank accounts 24/7
19:54 within minutes.
19:56 I can transfer to anybody else,
19:58 so there is an interoperability.
19:59 Second,
20:00 ease of access.
20:02 Third,
20:02 there was,
20:03 it's a,
20:04 completely
20:05 free.
20:06 So when I transfer money it's free.
20:08 When I scan to a merchant,
20:10 it's free for the merchant.
20:11 So it's highly subsidized
20:13 digital public infrastructure.
20:15 Customer or merchant,
20:16 nobody pays anything as of now.
20:19 And second,
20:20 it's
20:20 I can do it 24/7,
20:22 OK.
20:24 OK,
20:25 so then why this is a good scenario for developing a credit access.
20:30 The Basic idea was,
20:32 let's say I'm a roadside Kiana store or I'm a kiosk
20:35 or I'm a new to credit guy.
20:37 I have just a bank account.
20:38 If all my transactions happen in cash,
20:41 and then there's no digital footprint.
20:43 Now through this UPI and
20:45 managing through the bank and these payment apps,
20:48 I'm creating a lot of digital footprint,
20:51 so any lender,
20:52 be it bank,
20:52 be it fintech,
20:53 they can actually develop a cash flow
20:56 based lending based on my transaction history.
20:59 It can be 1 year,
21:00 2 years,
21:00 6 months,
21:02 which was different than credit score-based lending,
21:05 which was more liability side of lending.
21:07 So,
21:08 and,
21:08 and this cash flow based lending is a great
21:10 thing as long as you are looking for,
21:12 let's say a small ticket item,
21:13 ₹50,000
21:14 1000,
21:15 20,000.
21:16 Obviously if you are looking for ₹50 lakh of loan,
21:19 then this cash flow based lending is too risky.
21:21 All right.
21:23 So,
21:24 digital revolution of UPI,
21:26 which came in November 2016,
21:29 has gone up exponentially,
21:31 as you can,
21:32 as you can see here,
21:33 right?
21:34 OK.
21:35 So,
21:35 here are some basic statistics before I move forward.
21:38 So we have about 430 million unique UPI account.
21:42 Now India has about 952 adult populations,
21:46 so you can think about almost close to 50% of the adult population
21:50 has a UPI account.
21:52 What is the flow?
21:53 There is about 10 billion transactions happening per month,
21:56 amounting to $20.3 billion of transaction.
22:01 more than 75% of all retail digital transactions are on UPI now,
22:06 and there are about 70 million merchants
22:09 and 256 million QR codes available.
22:13 Then
22:13 on the mobile phone,
22:14 we have about 1 billion smartphones,
22:17 out of which 738 million smartphone users.
22:20 That means some people have more than one smartphone.
22:23 As of 2023,
22:25 average Indian mobile user consumption is 24.1 GB of data per month.
22:30 That's a lot.
22:31 Now I'm not saying
22:32 a
22:33 lot of this 24.1 goes in UPI,
22:36 but this just tells that India is
22:39 consuming data a lot.
22:42 And most importantly,
22:43 which is very,
22:44 very important for our,
22:45 our project,
22:46 is that by July 2016,
22:49 at least one member in every household in India
22:52 has a bank account through this Prime Minister Janan
22:56 data.
22:57 OK.
22:58 So,
22:58 as I said,
22:59 we collected the massive amounts of proprietary
23:01 data from 6 different sources for this.
23:04 So let me uh quickly uh go to the results because we don't have much time.
23:09 So this is a data this is a plot which I usually like to do without any fixed effects,
23:14 any control.
23:16 This is just raw plot,
23:17 which is state level aggregated month-wise
23:20 UPI transaction on the x axis
23:23 and the credit data from the TransUnion symbol,
23:25 and it says yes indeed there's a very high correlation
23:29 between UPI transaction volume
23:31 and credit expansion.
23:33 So this is a comforting graph,
23:35 which means that there is some signal in it which we want.
23:39 Here are some descriptions from the lenderwise trend.
23:41 So as you can see that
23:43 Fintech base is very,
23:45 very small,
23:46 and it started picking up around 2016 or 2017.
23:49 Now you will see there are two vertical lines there.
23:52 The first dotted vertical line is when UBI came in,
23:55 which is November 2016.
23:58 The second vertical line is September 2017,
24:01 which is when RBI
24:02 empowered UPI more by having a multi-bank account PSB model,
24:07 which means I can have as many bank accounts I want and I can
24:10 operate between bank accounts.
24:14 Here is my borrower type,
24:16 so you can see both prime,
24:17 subprime and more comforting than new to credit
24:20 borrowing going up.
24:22 OK.
24:24 Here's a loan composition by credit score.
24:26 These are all descriptives.
24:27 If you look at the fintech,
24:28 which is more interesting,
24:30 the new to credit and subprime growth around 2019 is much more,
24:35 right?
24:36 For example,
24:36 if you look at the blue line as well as the,
24:39 uh,
24:39 you know,
24:40 gray blue line in the bottom,
24:42 the jump is,
24:43 is much higher.
24:45 OK,
24:47 so
24:48 here's an identification strategy we have.
24:51 So what we said is the government of India came up
24:53 with a list of banks which they claim as early adopter,
24:57 meaning when UPI came,
24:58 those banks onboarded UPI much faster
25:01 and the late adopter who
25:03 onboarded later.
25:04 Now who are these banks?
25:07 Basically these are large banks which were technologically more.
25:13 Advance to adopt UPI much faster,
25:16 but in India,
25:19 large bank does not necessarily mean they are always in the very urban,
25:22 very income-oriented region because
25:25 major public sector banks are also very big.
25:28 So these large banks were all over,
25:30 both in rural as well as urban areas.
25:32 So what is the identification we create?
25:36 We said,
25:37 OK,
25:37 let's define something called UPI exposure,
25:40 which is nothing but the
25:41 amount of deposit in a zip code.
25:44 So
25:44 imagine
25:45 that
25:46 there are 5 banks in a zip code,
25:48 5 bank branches,
25:49 out of which 2 banks are the early adopters
25:52 of UPI,
25:53 and 3 banks are not.
25:55 Then you go back in that zip code and take the total amount of deposit as of 2015.
26:02 Of those two bank branches,
26:04 divided it by the total deposit of all the five bank branches,
26:08 and that gives me an UPI exposure.
26:10 Now what does this UPI exposure conceptually does?
26:13 It basically says that if a region has a lot more deposit
26:16 and not many banks adopt it,
26:18 then the UPI transaction becomes much faster
26:21 and much higher.
26:22 So we are developing an identification
26:25 through the UPI exposure.
26:26 And there is an enormous amount of variation,
26:28 as you can see on the UP exposure,
26:30 unlike concentrated in one region.
26:35 Here is the pre and post as of November 2016 when UPI came in.
26:40 If you look at the DID at the last column,
26:43 there's a number of loans in thousands.
26:46 For the FinTech,
26:47 the pre-base is very low,
26:49 so the jump post is pretty high,
26:51 but as I said,
26:52 this is not restricted just to FinTech.
26:55 This open banking spillover effect is also doing some good to the banks
26:59 and in BFCs.
27:01 Now we come to the first thing,
27:03 we just see that if the if the UPI exposure we have developed as identification
27:08 is
27:08 well correlated with the UPI
27:10 volume
27:11 and transactions,
27:13 OK?
27:13 And we see that it is actually,
27:16 uh,
27:16 you know,
27:16 highly correlated.
27:17 So from now on,
27:18 we are going to use UPI exposure
27:21 as our main identification
27:23 and not the UPI transaction
27:24 or volume.
27:26 OK.
27:27 So here is the parallel trend on the temporal dynamics,
27:29 if we have all the credit.
27:31 Then,
27:32 uh,
27:32 it,
27:33 as you can see,
27:33 it's kind of,
27:34 uh,
27:34 you know,
27:35 passed through the parallel trend assumption.
27:37 Uh,
27:37 the two vertical line has the same interpretation as I explained before.
27:41 This is for the fintechs,
27:43 uh,
27:43 post 2017 or a little bit before the growth is exponential in the FinTech.
27:49 Same with the NBFC
27:50 and same with the banks.
27:52 Now the regression results.
27:54 So this is the empirical specification we have,
27:56 and if we look at this empirical specification.
27:59 Then this is the first result which is impact on all
28:04 credit in the sense subprime prime,
28:06 neutral credit,
28:06 all credits.
28:07 What you find is
28:09 the region
28:10 which had a high UPI exposure,
28:12 meaning the deposit was much more in that zip code.
28:15 Post UPI coming in,
28:17 there is a growth.
28:18 In the amount of loan given
28:20 and the number of accounts
28:21 having the loan,
28:22 so there's an intensive margin
28:24 and there's an extensive margin.
28:26 Although I must say that we don't,
28:28 we cannot except the new to credit,
28:30 we cannot say whether a person
28:32 has only 1,
28:33 whether 5 persons has 5 loans or 1 person has 5 loans.
28:37 So it's a little
28:39 misconstrued on the extensive margin side,
28:41 but nonetheless,
28:43 the result still holds.
28:46 We see growth in subprime,
28:49 uh,
28:49 where there's 17% differential increase in the subprime credit value,
28:54 and we also see,
28:55 uh,
28:55 very comforting that is impact on new to credit.
28:59 So
29:00 total effect,
29:01 which is here we can actually taste extensive margin,
29:04 2% differential increase in new to credit in high exposure areas
29:08 and 14% differential increase in new to credit
29:11 in volume.
29:13 Um,
29:14 Now,
29:15 uh,
29:15 in the prime sector also it is increasing.
29:18 So that's something we thought about that why the prime is also increasing,
29:22 why are there
29:23 more credit constraints?
29:25 Looks like
29:25 if you look at the amount,
29:27 which is like about 6 million,
29:28 and the number of accounts,
29:29 which is like 61,
29:32 this is a number of accounts in thousands,
29:33 61,000.
29:35 Then the
29:36 ticket size for this prime sector is very small,
29:39 right?
29:40 So,
29:40 so there have been papers which show,
29:42 so first of all,
29:43 it's an average loan size is small.
29:45 Now if the average loan size is small,
29:48 then there is literature which shows that the prime borrowers
29:50 will actually go to the fintech than to the bank.
29:53 Why?
29:54 Because for a smaller amount,
29:56 the bank hassle of 1 day,
29:57 2 day paperwork
29:59 they want to avoid,
30:00 and the fintech attract those prime borrowers because of you know,
30:03 low cost and high convenience,
30:04 because I upload documents in the morning,
30:06 I get the money in the evening,
30:08 right?
30:08 So that is
30:10 one thing which we think is,
30:12 is,
30:12 is happening here.
30:14 OK,
30:15 now this is across the borrower level,
30:17 now across the financial intermediaries
30:20 who are leading this,
30:21 or everybody is getting benefited out of it.
30:25 So this is the increase in credit access across financial intermediaries,
30:29 as you can see there is a 217% increase in the number of loans through FinTech.
30:35 Compared to 38% and 21%
30:38 in terms of account by banks and NBFC
30:41 and in terms of volume,
30:43 83% increase by the FinTech compared to 27% and 23% increase in bank and NBFC.
30:49 So
30:50 as a base we see that everybody's getting benefited,
30:53 but the major chunk of benefit is taken away by the fintech,
30:56 although those are
30:57 majorly towards the,
30:59 as we'll see in a minute,
31:00 is new to credit
31:03 and subprime borrowers and small ticket loans.
31:06 So this is the impact of FinTech on all,
31:09 all borrowers,
31:11 where we found that nearly 84% increase in credit value
31:15 and 216 times increase in credit volume.
31:18 Now among this fintech,
31:20 what happened to the subprime?
31:22 The same thing,
31:23 there is a similar increase within the fintech
31:25 in subprime.
31:26 That means this because of this UPI,
31:29 the fintechs now are more confident
31:31 to give credit to the subprime borrowers
31:33 and
31:34 a similar increase in new to credit.
31:37 So we see that in the high exposure region post UPI coming in,
31:41 uh,
31:41 FinTech is actually giving more credit to the new to credit people,
31:46 which is basically the financial inclusion story
31:49 uh we want to tell,
31:50 and we see some increase in pride.
31:52 OK,
31:53 now,
31:53 what are the potential channel we,
31:55 we can think of uh for this?
31:58 There are two potential challenges we can think of in
32:01 the sense one is probably the regions which were unbanked before
32:05 but became
32:06 or came up with a banking account post 2014.
32:10 They would be more hungry for credit,
32:12 but their credit demand may not be that high.
32:15 So we think that the growth is probably happening through that,
32:17 which is
32:19 take us back to the story of
32:20 just having a bank account is not enough.
32:22 You have to tie it up
32:24 with some amount of digital payment history to develop a cash flow based lending.
32:28 And second
32:29 is
32:30 you need to have a very cheap
32:31 but very good internet connection for the UPI transactions to happen,
32:35 because
32:36 if,
32:37 imagine the cost of internet is high,
32:39 I really don't want to do it.
32:41 And second,
32:41 even if the cost is low,
32:43 if it is not good quality,
32:45 I open my bank account,
32:46 I want
32:47 to transfer,
32:47 and then it kind of hangs on,
32:49 then I will get,
32:50 I'll not trust it anymore,
32:51 right?
32:52 So these are very important.
32:54 And in India,
32:54 what happens in 2016,
32:56 a little bit after UPI came,
32:59 Reliance Geo came in
33:00 and kind of,
33:01 uh,
33:02 you know,
33:03 made the internet very,
33:04 very cheap.
33:05 I'll show you in a minute how cheap that is,
33:07 and we think the regions where this Reliance Geo came in,
33:11 that's where the UPI growth becomes max,
33:13 and,
33:14 and third,
33:15 we do our digital verifiability of our claim
33:18 from a FinTech player.
33:20 OK.
33:21 So on the
33:22 JDI account,
33:23 we do the triple difference and what we found is uh
33:27 across
33:28 all credit.
33:29 Now when I say all credit is subprime prime mutual credit altogether,
33:33 but we found the last row,
33:35 high GDI times post,
33:36 basically saying that
33:38 if a region zip code
33:40 has a very high number of GDY accounts,
33:42 then UPI coming in
33:44 increases the amount of loan
33:46 by 34
33:48 million,
33:48 right?
33:49 Now if
33:51 The region is already a high exposure region,
33:54 and that has high JDI account,
33:57 then post UPI
33:59 that differential increase is even much more,
34:01 so the total increase will be 34 + 7,
34:03 right?
34:05 And similarly found in the number of accounts.
34:09 Then we look at
34:11 uh
34:12 You know,
34:13 we look at the FinTech credit volume,
34:15 and we found that it is actually a lot happening on the
34:19 on the fintech side.
34:20 Similarly,
34:21 within FinTech,
34:22 we look at the new to credit
34:24 and we found that the region,
34:26 which has a higher number of JDI accounts
34:28 and a higher deposit base,
34:30 the post UPI,
34:32 the new to credit getting
34:33 a lot of loans,
34:34 which is kind of the hypothesis
34:36 we had.
34:38 Now what happens to the
34:40 internet connection?
34:41 Now,
34:42 Reliance Geo actually rolled out 4G towers in September 2016,
34:46 and they started the service from December 2016.
34:49 From December 2016 till March 2016,
34:52 they gave
34:54 almost zero cost Wi Fi,
34:56 very high quality.
34:58 Whereas back before Geo came in,
35:00 the cost of 1 GB of data per month
35:03 was almost like $5
35:05 and 4eo it came down to 50 cents.
35:08 And then after March 2016,
35:10 they continued with a very cheap internet rate.
35:13 How cheap it is,
35:14 you can see here.
35:15 So this is the reduction
35:17 in cost of 1 GB data in India.
35:22 And you can see that the
35:23 exponential curve inflection point is around 2016.
35:28 Right?
35:29 So we use this,
35:30 we have this telecom tower location data,
35:33 where we know every telecom tower location of India,
35:35 the location,
35:36 whether there's 3G,
35:37 4G,
35:37 and their provider.
35:39 So we,
35:40 our hypothesis is the region where the
35:42 geo.
35:44 is present with their 4G
35:46 there because I can do a lot more
35:49 without interruption
35:50 and high quality transfer
35:52 of money.
35:53 There I'm developing a better
35:56 digital payment history,
35:57 so the loan will be better.
35:59 So again,
35:59 we
36:00 use this proximity as a measure and triple difference,
36:03 and what we found is
36:05 across the board
36:06 there is an increase in high exposure pin code with
36:09 geo towers.
36:11 But the increase is huge in terms of volume
36:14 by
36:14 fintech lenders,
36:16 OK?
36:17 Banks and NBFCs are not that high.
36:20 Now,
36:20 this is a proximity to geo tower,
36:22 so if a customer is in a high exposure area and
36:25 there is a proximity to the geo tower is low,
36:28 then
36:30 there is a FinTech subsample
36:32 from the FinTech subsample,
36:33 the fintechs are giving more loans in terms of number of accounts
36:36 as well as amount.
36:38 The same thing for new to credit,
36:40 but what is also interesting is that the counterfactual.
36:44 So you also try to look at if you look at the last row and the second row,
36:48 we try to do the same thing for non-geo towers which are also giving 4G,
36:53 but the cost of data was high,
36:56 and we see that there
36:57 nothing is actually happening.
36:59 So that means a cheap internet
37:01 and a quality internet is very,
37:02 very important for this trade exclusion to
37:05 continue.
37:06 OK.
37:07 Uh,
37:07 I guess I have what,
37:08 about 5 minutes?
37:11 5 minutes.
37:12 Perfect.
37:13 OK,
37:13 so now
37:14 we look at a fintech provider,
37:17 uh,
37:18 which is
37:19 basically
37:20 does.
37:21 So what they do is they will all the Kiana store roadside kiosk,
37:25 they will go and put their QR code free of cost,
37:28 and they say that,
37:29 you know,
37:30 you ask your customer
37:32 to pay through the QR code
37:34 that way I will have data on what is your transaction flow,
37:38 your revenue flow,
37:39 and based on that,
37:41 once you build that for 4 to 5 or 6 months,
37:43 I'll be able to give you a loan.
37:45 So this is what this fintech company does,
37:48 right?
37:49 So what we wanted to do is we wanted to show that what is happening
37:54 for one individual level,
37:55 because so far everything we have done at the
37:57 zip code level.
37:59 So in this case,
38:00 what we found is that
38:03 The UPI exposure measures are not correlated with the internal credit score.
38:08 However,
38:09 UPI exposure measure is correlated to the size of the loan given by the company.
38:14 Now they are also correlated,
38:16 you know,
38:17 with the law QR transaction and the law of QR transaction count.
38:21 So what this table basically says that
38:23 if
38:24 a merchant
38:26 has high QR transaction
38:28 and number of transactions is high,
38:30 then definitely the amount of loan approval and the amount of loan
38:35 loan size goes up.
38:37 But it also says
38:39 that if
38:40 I
38:41 am in a region of high exposure,
38:43 high UPI exposure.
38:45 Then
38:46 there is a network effect.
38:48 Then that region
38:49 also growing,
38:50 so that's why you get more loans.
38:52 So to tease this out more,
38:55 we actually,
38:55 in the second result,
38:56 we do an interaction,
38:58 and this interaction,
38:59 what you say I see that.
39:01 In a high UPA exposure area,
39:03 I can have two merchants.
39:05 One merchant
39:06 whose
39:07 own QR code transaction and volume are low.
39:11 Another merchant where his own QR code transaction volume is high.
39:15 And we see that in the high exposure area,
39:17 the guys who are doing more transactions own
39:20 their own transaction more
39:22 on the UPI,
39:24 they are getting a much better and a bigger loan,
39:26 which shows that
39:27 it is not only important
39:29 that you do a lot on the UPI transaction from the merchant side,
39:32 you should also be in a region
39:34 where there's overall growth and thus more banks
39:37 and BFCs and Fintech gifts loans there,
39:39 so there is a network
39:40 externality effect.
39:42 OK.
39:43 So whatever we found this important feature is that the
39:46 bank accounts with digital payment history is a powerful combination
39:50 for credit exclusion.
39:55 India's JDY drive
39:57 helped every household to have a bank account,
39:59 and that was very,
40:00 very necessity.
40:02 UPI
40:04 and having an open banking framework which is customer permission data sharing
40:08 is at the heart of this digital revolution in India,
40:11 and
40:12 what this UPI through open banking does for India is basically
40:16 customers
40:17 can create
40:19 real-time,
40:20 zero cost,
40:21 digitally verifiable financial transaction history,
40:24 and they can share to any financial intermediaries
40:27 to
40:28 help them
40:29 to generate
40:30 a cash flow-based lending model.
40:33 In the conclusion,
40:34 I think
40:35 We have tried to answer the first order question,
40:38 which is
40:39 how do you do more credit expansion,
40:42 especially to the financial inclusion,
40:44 and we think
40:45 this digital public infrastructure
40:48 is very,
40:48 very important,
40:49 and we introduced a new channel through that
40:52 and we document
40:53 that there is an increase in credit access because of this revolution in India
40:58 to the marginal.
40:59 And the unserved borrowers,
41:01 so we call it the FinTech inclusion too
41:04 in a way,
41:04 and we think
41:06 you know
41:07 this
41:07 account aggregator and some other initiative government
41:10 of India is having will further enhance
41:12 this financial
41:13 inclusion.
41:15 OK,
41:16 I guess I'm on time,
41:17 I believe.
41:17 Thank you so much,
41:18 Pulak.
41:18 That was fabulous.
41:20 I think,
41:20 uh,
41:21 everyone,
41:21 uh,
41:21 is waiting to ask questions.
41:23 So why don't I open it up?
41:25 Sure,
41:25 and
41:26 colleagues,
41:26 uh,
41:27 raise your hand,
41:28 and I'll turn to you,
41:29 right?
41:29 So.
41:37 So as others are asking or thinking about questions,
41:40 Pulak,
41:40 can you reflect a little bit about the rollout of GO
41:44 and
41:45 the interaction of speed
41:47 and
41:48 low price and kind of enabling,
41:50 uh,
41:51 the,
41:52 the access to credit that you described?
41:54 How,
41:54 how much it varies within India.
41:57 So,
41:57 so basically what happens is,
41:59 let's,
42:00 so the geo revolution happened in December 2016.
42:03 So imagine a scenario before that.
42:06 Before that,
42:07 1 GB of data
42:08 in a month used to cost $5.
42:12 Now,
42:12 if you look at these people for which we are trying to do the financial inclusion,
42:17 this is a pretty costly,
42:18 uh,
42:19 you know,
42:19 data acquisition,
42:20 right?
42:20 $5 is too much of money for them,
42:22 uh,
42:23 and that was also just 1 GB,
42:25 uh,
42:25 in a,
42:25 in a month.
42:26 Uh,
42:26 so then Geo came and Geo
42:29 reduced the price to almost like 60 cents
42:31 in the,
42:32 in,
42:32 in,
42:32 and that also they said
42:34 for 90 days,
42:35 not for one month,
42:36 right?
42:37 So all of a sudden what happens in the rural sector and in the semi-urban sector.
42:41 The,
42:42 the smartphone penetration with internet gone up exponentially.
42:47 And then they started surfing,
42:49 they started doing all sorts of things,
42:50 and obviously the UPI also becomes easier for them to transact on UPI because
42:55 remember,
42:56 UPI is moving off your transaction,
42:59 uh,
42:59 you know,
42:59 so it,
43:00 it is going to consume some amount of data.
43:02 And because it's going to consume some amount of data,
43:05 for a person like you and me,
43:07 it's different.
43:08 We don't really care that much,
43:09 but for the semi-urban,
43:11 rural,
43:12 you know,
43:12 tier 2,
43:12 tier 3 rural areas,
43:14 there are two main issues one,
43:17 Is this going to cost me too much?
43:19 If you say that it's not going to cost me at all,
43:21 then the second question is,
43:23 OK,
43:23 imagine a scenario.
43:25 I opened my GPay
43:27 and I scan the QR code,
43:29 and I put the money and I hit the button to
43:31 money go and all of a sudden the internet is gone.
43:34 Then I worry,
43:35 is my money safe,
43:37 and you know there is no way of figuring that out until the internet comes back,
43:40 and it may happen that my money,
43:43 you know,
43:43 gone somewhere else.
43:44 So that trust is also taken care of by
43:47 the quality of the internet.
43:49 So Geo did two things one is reduce the price
43:52 drastically.
43:54 And that made over time the
43:57 other players like Airtel,
43:58 Vodafone,
43:59 BSNL,
43:59 they also reduced the price because
44:01 the main leader has reduced the price,
44:03 so they have to win the business.
44:04 Second,
44:05 because Geo's quality of internet was very good,
44:08 the other players saw that.
44:10 A lot of people are moving to Geo
44:12 now.
44:12 In order to stop that,
44:13 they also started giving a better service.
44:16 So I think that really changed the UPI adoption,
44:20 although it has other parameters like free,
44:22 real-time,
44:23 everything,
44:24 but this really changed the game for the UPI adoption.
44:30 And what is particular about uh
44:34 India's system
44:35 in terms of interoperability,
44:37 because when you look at other countries,
44:39 it's not very clear
44:41 if the systems are interoperable
44:45 and uh come on with your question and let's take
44:46 a couple of more and turn back to Pollock then.
44:49 OK,
44:50 so to answer your question,
44:52 we're going to ask a question as well and Pollock,
44:54 then you can kind of take them together,
44:57 come on in.
45:05 We can't hear you.
45:06 Uh,
45:07 at least I can't,
45:07 so.
45:14 OK.
45:16 Sorry,
45:16 did you,
45:16 did you call on me?
45:17 I'm sorry,
45:17 Shami.
45:19 No,
45:19 Shavi,
45:19 if you want to please come on in because Maria had her hand raised as well.
45:23 But Shavi,
45:23 come on in with your question if you have one.
45:25 OK,
45:25 sorry,
45:26 sorry,
45:26 sorry.
45:26 Yeah,
45:26 so Pulak,
45:27 thank you so much for this very nice presentation.
45:29 I mean,
45:29 you know,
45:30 uh,
45:30 very,
45:30 very instructive.
45:32 So I,
45:32 I actually have,
45:33 uh,
45:33 questions all over.
45:34 So,
45:35 so I'm gonna try to,
45:36 uh,
45:36 number one,
45:39 Now that FinTechs uh can observe the cash flow through this open data,
45:44 um,
45:44 are the terms of the contracts
45:47 changing from the,
45:48 you know,
45:49 traditional microfinance or traditional credit
45:52 where,
45:52 you know,
45:52 it was a fixed amount every month.
45:55 Does that now,
45:56 do they now tailor the,
45:58 the,
45:58 the payments to the cash flows,
46:00 uh,
46:00 and so,
46:01 you know,
46:01 allowing some flexibility in how they,
46:04 how they repay.
46:04 That,
46:05 that's question number 1.
46:06 Question number 2,
46:07 I'm,
46:08 I'm really interested.
46:09 In,
46:10 how is the tax authority looking at all of this?
46:13 Because,
46:14 you know,
46:14 so before,
46:15 you know,
46:16 income taxes were,
46:18 I guess there were some threshold above which you had to pay,
46:21 below which
46:22 you didn't have to pay and that threshold was
46:23 determined by how expensive it was to do a
46:26 Uh,
46:27 an audit,
46:27 right?
46:27 I mean,
46:28 you had to go there and,
46:29 you know,
46:29 visit the premises and see,
46:31 you know,
46:32 how much money they were making.
46:33 Now,
46:33 I mean,
46:34 it's free,
46:34 right?
46:34 I mean,
46:35 you,
46:35 all you gotta do is look at the history of,
46:37 of
46:37 of,
46:38 of transactions,
46:39 and then,
46:39 you know,
46:40 you can,
46:40 you can get a,
46:41 a much better assessment of,
46:42 of what Eco might be
46:43 from just,
46:44 just,
46:44 you know,
46:44 you could be sitting in Bombay and just,
46:46 you know,
46:46 just collecting this data,
46:47 right?
46:47 So,
46:48 so again,
46:49 what,
46:49 what's the,
46:50 what's the tax authority thinking about,
46:52 about this.
46:53 And,
46:54 and then,
46:55 you know,
46:55 our people responding to,
46:57 to this potential threat of,
46:58 of,
46:59 of taxation.
47:00 By,
47:01 by not using UPI,
47:02 right?
47:02 I mean,
47:02 are people close to this threshold
47:05 basically relying more on cash than on,
47:07 than on UPI on,
47:09 on,
47:09 on the digital payment
47:10 just because they're,
47:11 they're afraid,
47:12 right?
47:12 Or,
47:12 or I guess more broadly,
47:14 are there segments of the population that are,
47:16 that are reluctant to use UPI,
47:18 uh,
47:19 and,
47:19 and why that might be the case?
47:20 Uh,
47:21 I'll,
47:21 I'll stop here.
47:21 So let me,
47:22 let me take your first question.
47:24 So yeah,
47:24 fintechs are coming up with more innovative product.
47:27 Which is more,
47:27 uh,
47:28 so if you look at the
47:29 uh retail space of lending,
47:30 they have products starting from 15 days loan to 180 days loan
47:34 with different,
47:35 uh,
47:36 you know,
47:36 interest rates.
47:37 So they are,
47:37 they're way more flexible
47:39 in,
47:39 in terms of this.
47:40 If you look at the micro SMEs,
47:42 like for example,
47:42 this fintech company I told you about,
47:45 they have,
47:45 uh,
47:46 you know,
47:46 settlements in the lineup,
47:47 they give the line of credit.
47:49 So you just to get a line of credit,
47:50 you can increase it based on the payment structure and the
47:53 And,
47:54 and
47:54 their payment is the end of the day like microfinance.
47:56 So basically what happens is
47:58 I know every transaction,
48:00 it's getting your bank account at the end of the
48:03 day,
48:03 let's say around 8 p.m.
48:05 I just take,
48:06 you had a cash flow of let's say $100
48:09 and you were supposed to pay me $5.
48:10 I just take directly $5 from your bank account.
48:13 So this kind of very flexible structuring is happening.
48:16 On the taxation thing,
48:17 it's not really a tax threat for anybody,
48:21 particularly for the reason that if it is people like you and me,
48:24 we are transacting based on our deposit coming from our salary or other income,
48:29 which is either way we have to disclose.
48:30 We are not adding or deleting anymore.
48:32 On the margin side,
48:34 because of GST,
48:36 you are already,
48:36 it doesn't matter how small,
48:38 how big you are,
48:39 you have to file a GST.
48:41 So because you are filing a GST either way,
48:43 whether you are transacting on the UPI getting on your thing,
48:47 that is completely separate.
48:48 So tax authority at this moment doesn't have
48:52 a hand in the UPI transaction.
48:54 That's completely separate to NPCI.
48:56 If the tax authority wants somebody's UPI transaction,
48:59 they have to route it through the Supreme Court of India,
49:02 so they cannot directly get that.
49:04 But the merchants,
49:05 when I talked to in another paper which we are doing,
49:08 uh,
49:08 the merchant they talked to,
49:09 they say that initially
49:10 they were more afraid not for UPI for the GST,
49:14 because the moment you put the GST,
49:15 you are basically pretty much telling everything about it.
49:18 But later on they figure out that if
49:19 I actually disclose the entire revenue through GST.
49:23 Then that's another digital transaction process through which I can get loan.
49:28 And that's why you will see that India's GST,
49:30 uh,
49:31 you know,
49:32 is becoming more and more every quarter,
49:34 because more and more small micro SMEs are coming on board.
49:38 Yeah.
49:40 Thanks,
49:40 Bullock.
49:40 Uh,
49:41 Maria and Emilia,
49:42 why don't you guys come in?
49:44 No,
49:44 thank you very much.
49:45 This was really a fascinating,
49:46 uh,
49:47 presentation,
49:48 and I have a lot of questions but I will restret
49:50 for a reason of time mine.
49:52 One would be,
49:53 is anything that we can say on the distributional impact
49:57 that is,
49:57 you know,
49:58 like,
49:59 perhaps,
49:59 uh,
50:00 consumers or lenders,
50:01 do you have any information
50:02 on the income or
50:04 Whether it really allows even,
50:07 you know,
50:07 the poorest or relatively poorest people to,
50:10 to access.
50:11 And second,
50:11 what
50:12 would be the cost of subsidizing cost?
50:15 Both,
50:15 you know,
50:15 you mentioned
50:16 that the internet has been highly subsidized
50:19 and there've been lots of subsidization of infrastructure
50:23 more generally.
50:25 So thank you very much.
50:26 Yeah,
50:27 so the first question,
50:28 distribution,
50:28 I have,
50:28 I have shown that uh chart.
50:30 Let me see if I can uh pull it out again.
50:32 Uh,
50:33 basically,
50:35 It's uh,
50:36 hold on.
50:42 So we don't know the income,
50:43 but we have the cut based on prime,
50:45 super prime,
50:46 subprime,
50:46 new to credit.
50:48 Um,
50:49 that should give some idea about the
50:51 uh distribution.
50:53 Can you see this graph?
50:55 Uh,
50:56 so we can assume they're new to credit either they're a very young population
50:59 coming out of college or they are actually lower part of the income,
51:03 income strata,
51:05 but,
51:05 but the expansion is happening all across,
51:07 not just one group.
51:08 So the distribution is increasing
51:10 both in prime but majorly happening in new to credit and subprime,
51:15 uh,
51:15 to answer your first question.
51:16 Uh,
51:16 second question is
51:18 nobody is subsidizing the internet.
51:20 So the Reliance is a private player.
51:22 It's their business model.
51:23 Their business model is to have
51:26 more customers taking their internet and thereby because of
51:30 scale they can reduce the and the technology advancement.
51:34 So they're bringing more competition basically.
51:36 So the government of India is not subsidizing the internet at all.
51:38 It's a private player play.
51:41 Yeah.
51:43 Pulak,
51:43 uh,
51:44 how do we know that reliance isn't using one part
51:47 of its business to subsidize another part of its business?
51:50 No,
51:50 within the business they can,
51:51 but I'm saying government is not getting into this,
51:53 and most likely they are doing that.
51:55 They are substituting because
51:57 through this more people coming on board,
51:59 they get more business out of advertisement and
52:01 several other things because it's a conglomerate,
52:03 right?
52:03 So.
52:05 Uh,
52:05 Emilio,
52:07 And thank you very much for this uh presentation.
52:09 It's
52:10 very interesting and
52:12 Hypothesis is very relevant for a lot of the work that we do.
52:15 Um,
52:16 my question is more on,
52:18 uh,
52:18 any reflections from you on how this impact,
52:22 the mechanisms that led to these impacts of your
52:25 part of your argument is
52:27 the establishment of UPI and also the open finance framework.
52:31 However,
52:32 as far as,
52:33 as,
52:33 as we've been able to exchange with some organizations
52:36 leading the open finance framework like Sahamati and,
52:38 and the comments from RBI is that there has been a.
52:42 The,
52:42 you know,
52:42 the,
52:43 the open framework and paper was established,
52:45 but the process of implementing in practice and actual
52:48 sharing of data has been very slow because,
52:52 um,
52:53 the actors,
52:54 um,
52:54 connecting to the open finance framework has been,
52:57 have been,
52:58 uh,
52:58 slowly pouring,
52:59 especially the big actors like the state banks,
53:01 right?
53:02 Um,
53:03 and so,
53:04 I would have imagined,
53:06 uh,
53:06 or expected kind of a,
53:08 a lag effect,
53:09 but your,
53:10 your effect.
53:11 is,
53:11 is pretty strong and pretty evident,
53:13 so I wonder
53:14 how much of that is due to open finance,
53:16 how much of that is due to plain UPI.
53:19 Um,
53:20 and,
53:21 uh,
53:21 and whether you have any updates in terms
53:23 of the actual implementation process of API with
53:27 fewer actors participating initially yet,
53:30 and now still,
53:31 a lot of actors are still pending,
53:32 you know,
53:33 the actual,
53:33 uh,
53:34 implementation of OPI
53:35 transactions.
53:36 OK.
53:37 So,
53:37 I think there is a confusion here.
53:39 So when,
53:39 when I said the open banking,
53:41 open infrastructure is UPI only.
53:44 And not the account aggregator.
53:45 I think you are talking about more of an account aggregator.
53:48 So account aggregator is a very recent phenomenon
53:51 and it's a much broad.
53:53 It kind of tells you not only your payment history,
53:56 your investment,
53:57 your insurance,
53:58 pretty much everything.
53:59 That's taking
54:00 a slow progress mainly because there is not just one actor.
54:03 There is the banking industry has to come in.
54:05 Insurance industry has to come in.
54:07 Brokerage houses have to come in.
54:09 So that's taking a lot of time.
54:11 But
54:12 that's the second stage of the open banking,
54:15 I guess.
54:15 I see,
54:16 I see.
54:17 The first stage is just the UPI.
54:18 So what I meant is completely the UPI,
54:21 which has no issue for the banks
54:24 and the third party player to come in.
54:31 Thanks for clarifying.
54:31 Yeah,
54:32 that's useful.
54:35 Thanks,
54:35 Pulak.
54:36 We got
54:36 room for one more question if someone has one more question.
54:44 If not,
54:44 uh,
54:45 Pulla,
54:45 maybe you wanna,
54:46 uh,
54:46 give some closing thoughts on,
54:48 uh,
54:49 the research that you're doing and,
54:51 uh,
54:52 some of the other research programs that you have.
54:55 Sure,
54:55 so basically this is my 3rd paper in this,
54:58 and this is a much broader picture,
54:59 but I think
55:01 India is a path of digital revolution.
55:04 They have done a lot,
55:06 and in the heart of it,
55:07 I think is the democratization of it.
55:09 So there is no monopoly,
55:11 uh,
55:11 you know,
55:11 uh,
55:11 basically most of this government of India or RBI or NPCI done it
55:16 and allows any third player to build on it.
55:19 I think that's the major difference between.
55:22 Uh,
55:25 any other DPI across the world than India,
55:26 and with the account aggregator,
55:28 if it is successful,
55:29 I'm hoping in a year or two it will become,
55:31 uh,
55:32 then that will,
55:33 uh,
55:33 you know,
55:34 give a huge push,
55:35 not only to financial inclusion,
55:37 but also to the default,
55:39 because at this moment what is happening is that
55:42 because these are small ticket loans,
55:44 and I can,
55:45 if I and you both have a screening technology through which
55:49 I can develop a risk model based on your cash flow.
55:53 If I'm a prime borrower,
55:54 I can apply to 5 Fintech lenders,
55:56 and all the 5 within an hour will provide me the credit,
56:00 so I'm getting extremely leveraged.
56:02 So right now there is the mechanism because these are happening so fast.
56:07 There is no mechanism to figure out that I'm actually at the same time looking for 5
56:12 loans,
56:12 and that increases the default
56:15 chance of default.
56:16 So I think the account aggregator coming in and if they
56:19 kind of update the data every 5 minutes or an hour,
56:23 then
56:24 in 1,
56:26 the lenders will have a lot more.
56:28 Information to develop a better product for you.
56:30 And second,
56:31 this overleverage can be reduced.
56:34 RBI is dealing with it by expanding the weighted,
56:37 you know,
56:37 weight average,
56:38 uh,
56:38 but still I think the speed has to be somehow taken care of.
56:42 But I,
56:42 I guess I,
56:43 I will,
56:44 uh,
56:44 you know,
56:45 request all of you to look at India.
56:47 We are
56:47 going through a fascinating stage,
56:49 and,
56:50 you know,
56:50 so it's great things are happening.
56:54 Thanks so much,
56:55 Pulla for uh giving us this really fascinating talk.
56:58 I think,
56:58 um,
56:59 everyone
57:00 is keen to learn about the impact of India's digital revolution because there
57:04 has been a lot of talk on the extent of India's digital revolution.
57:09 Uh,
57:09 colleagues,
57:10 we will be making,
57:11 uh,
57:11 the slides available to everyone,
57:14 and there'll be a recording of the seminar that we'll be posting.
57:17 Thanks so much for joining
57:19 and thank you again for luck for,
57:20 for the presentation.
57:22 Thank you.
57:23 Great day.
- add-style
- lp-body-content