Customers without credit scores can take the digital journey to get loans
Mumbai-based CreditVidya, a fintech start-up, uses alternative data sources to assess fraud and risk. It has recently raised $5 million from Matrix Partners and had previously raised $2 million from Kalaari Capital. The money is being used for product development and hiring manpower. A lot of the investments are going into research and development and setting up the team right, which will include data scientists from the US. The plan is to have a total of 146 employees by end of 2018, said the founders of the company. In an interview with Mint Money, Rajiv Raj and Abhishek Agarwal, co-founders of CreditVidya, spoke about why fintech companies and peer-to-peer (P2P) lending platforms are looking at alternative data, and as a consumer how you should take care of your digital footprints. Edited excerpts:
Currently you work with over 20 banks and non-banking financial (NBFCs) that are looking to assess customers of small unsecured credit. What is the quality of these banks and NBFCs?
Rajiv Raj: We have a mix of small and big banks and NBFCs. We have big banks such as State Bank of India, ICICI Bank Ltd and Axis Bank Ltd. There are many micro services that these companies use. One can use full-stack services—from acquisition to underwriting—and there are also companies that use us only for email fraud or employee verification.
Abhishek Agarwal: We are also in talks with an MNC (multinational corporation) bank. Right now, 10 relationships are with large banks and NBFCs, out of the 27, and remaining are in mid- and small-sized banks. Every bank is focused on retail loans and in that pie on unsecured lending. Personal loans, consumer durables and two-wheeler loans are the segment where there has been a tremendous rise.
A lot of fintech companies that started as e-wallets and payment gateways are now moving to credit, obviously because margins are low for payments while credit has higher margins. Has any fintech company contacted you to help assess first-time borrowers?
Agarwal: We are under strict NDA (non-disclosure agreement) norms currently for fintech companies, so we can’t name them. But we are in talks with e-wallet, e-commerce and payment gateway providers. All of these companies want to give credit because of the low margins in their core product. These companies have become enablers of credit.
Since you are working with e-commerce companies, do you eventually see them coming out with credit products?
Agarwal: The e-commerce companies are going to move to a pre-approved model. They will be able to pre-approve you for say Rs50,000 credit for you to shop. And that pre-approved loan will be backed by one or more lenders.
I know of fintech companies that have tie-ups with NBFCs and don’t report credit behaviour to the credit bureaus. Is there a need to fear where the industry may get into a bad credit cycle because of not reporting bad loans?
Agarwal: The ones you are talking about are bullet payment products, which is basically cash advance. Theoretically, these are not loans and hence are not liable to report to a bureau. Also, the volume of the data is currently very low. The way they have structured this exercise is, while there are no strict penalties, the consumers will not be able to use the platform if they default. I don’t think it is too big a challenge. Volumes are too small right now.
Recently, P2P regulations came out. These companies will have to start reporting to credit bureaus. Has any P2P platform approached you to use alternate data?
Agarwal: We are currently working with three P2P lenders. Here again, it is for risk assessment of first-time borrowers. People who are digital savvy and want to access this facility, are first-time borrowers and under 35 years. Cibil’s (a credit bureau in India) penetration in the 25-35 age group is poor. Hence, 75-80% of the cases will have no Cibil score.
Raj: These are thin-file customers who don’t qualify for loans. You can’t actually assess a customer based on one consumer-durable loan. There is hardly anything there.
Agarwal: P2P companies are targeting digitally savvy customers who are young and not likely to be on Cibil.
Considering that more companies are going to use digital data to assess consumers, how should consumers take care of their data footprints?
Raj: Digital data tracking is based on consent. First, you need to educate customers that the information you are giving to a lender will be used to assess your risk for the loan. The awareness will come over a period of time. When you are assessing customers who don’t have a credit bureau footprint, one of the things that lenders says is that if you want to qualify for a loan, you have to go through the digital journey. I clearly see this for a particular segment—small ticket and unsecured lending. The moment such a customer will go to a point-of-sale and look for credit, they will check her credit bureau data and income document. If they don’t find it, they will ask her to download an app and give consent to use her digital footprints.
Why are the traditional credit bureaus not using alternative or digital data to assess customers?
Raj: One, there is a regulatory issue. Two, they have never done this before.
Agarwal: Experian (a credit bureau) in the US has been around for the last 40 years. Digital lending in the US exists for the last 12 years. Experian never used alternative data in the US. It is not in their DNA. All the traditional bureaus in India are heavily influenced by their parent companies in the US. There is no product that the bureaus have launched in India that is only for the Indian market. They haven’t done anything that is new and specific to India.
While analysing customers, what parameters do you use to evaluate credit worthiness?
Agarwal: You look at five types of fingerprints—social finger print (anything you put on social media), device fingerprint (such as SMS), browser fingerprint (anything that identifies your device), click stream fingerprint (how fast you type) and biometric fingerprint (the physical fingerprint). The type of data source we use can vary. For instance, from an expense perspective, how does your profile look? Do you spend your money as soon as you get the salary? Do you have savings? From transactions perspective, we have noticed that a good postpaid bill payer is also good with credit card financing. We also look at location profile. For instance, we have seen clusters of people in one small location that tend to be defaulters, indicating a higher probability of people in this neighbourhood to default. There is something called network size. Network would be connected to people, location or company. If two people are talking to each other, they are likely to be of the same socioeconomic strata. We can build a hypothesis around it. We also ask questions when they take Rs5 lakh loan. Are you earning Rs75,000 per month? When you are taking Rs40,000 or Rs50,000 loan, the question we ask is, if the minimum salary is more than Rs15,000 a month. We can know the average income of an individual from network data.
With traditional credit bureaus you can correct your credit history. Is that possible with alternative credit data too?
Raj: In a traditional credit bureau, the credit score of a customer is based on severity. Someone who has defaulted 3-5 years back but last 2 years the payment has been consistent, obviously the score will get better. This is the same structure for digital data as well. If you don’t see any derogatory information, the scores will be higher.