If you’re someone with a low income and no significant credit history, getting a loan can be tricky. You may have to postpone a certain purchase because no lender would approve a loan if your annual income doesn’t fit their financial prerequisites. However, the situation may have eased today due to massive revolution in the fintech space. In fact, now all you need is an android smartphone to get instant digital credit. TransUnion-CIBIL data for January to March quarter reveals that over four million ‘new to credit’ Indian customers have loan-related queries every month. These include consumer durable loans, personal loans, credit card loans, housing loans and automobile loans. Fintech companies, that hold a market share of 56% in the ‘new to credit’ segment, are using machine learning technology to assess the repayment ability of their customers with the help of their smartphone data. This is popularly known as smartphone footprint.
Smartphone footprint could refer to any information available on one’s smartphone. It could include SMSes regarding banking transactions, phonebook contacts, app usage behaviour, etc. Multiple fintech companies are aggressively using this model to get insights from the user’s smartphone data that can be used as inputs for predictive risk models used in credit underwriting.
Take example of FinBox, a fintech company from Bengaluru that provides tailored credit products to their customers and partners. It uses privacy-friendly and explicit user-consent driven flows to analyse behaviour data and non-personal SMS data to assess borrower credit worthiness.
“This data is used to construct ‘features’ or independent variables, which are then put through machine learning models to recognise patterns that differentiate between high risk and low risk customers. This process results in a ‘score’ that provides guidance in terms of a credit decision to the lender. This entire process is driven by user consent, and the access to data is use-case driven and time-bound. It cannot be sourced for one purpose and then used for another,” said Anant Deshpande, co-founder at FinBox.
This alternative data is almost always used in addition to formal data such as credit bureau check etc. but in many cases, formal data availability is minimal such as in the rural parts of the country or with micro entrepreneurs. In those cases, alternative data underwriting helps deepen financial inclusion by providing a strong proxy for one’s creditworthiness. Furthermore, in case of credit bureau errors, this data can be used to build a case for a borrower’s creditworthiness.
Data from these models is used to disburse loans of up to Rs 25,000, which constituted for 60% of total personal loan originations in Q4 2020, as per data from TransUnion CIBIL. For loans higher than Rs 25,000, most fintechs dig deep on information, i.e., a net banking login or upload bank statement.
At Finbox, Deshpande suggests, “We seek storage permission so that your KYC and other relevant documents can be securely downloaded, camera information permission to enable you to click photos of your KYC documents and others, and upload the same on the app during your loan application.”
The company collects/monitors only financial SMS sent by 6-digit alphanumeric senders from your inbox which helps us in identifying the various bank accounts that you may be holding, cash flow patterns, description and amount of the transactions undertaken by you. This information is compiled and analysed on an aggregated basis.
“We share this information with only an RBI-regulated NBFC that is meant to access this information to make a lending decision,” Deshpande added.
Research covering banks, credit unions, finance companies, and FinTechs has shown that over 80% incorporate alternate data in at least some part of their lending decision making. Of these, it was seen that FinTechs and other non-banks are most likely to use alternative data for every loan application. However, the usage of alternative data is picking up across the ecosystem not just for credit underwriting but for other use cases such as fixing insurance premiums etc. too.
“The key insight here is that privacy-driven regulations and frameworks for data sharing have allowed users to be more aware of their own data and utilize it to their advantage without compromising with their control over this information,” Deshpande pointed.
How does a fintech company make sure that the concept of smartphone footprint does not breach user privacy?
“At FinBox, we only access customer data based on clear, explicit, and transparent user consent. Users can control the kind of data accessed at all times. Our risk assessment products are built in alignment with DEPA (Data Empowerment and Protection Architecture), to ensure data is secure and encrypted both at rest and at motion,” Deshpande explained.
While creating an account on the platform requires basic information about a user (name, birthdate, email ID etc), we only collect the information needed for our specified purpose, after seeking the user’s permission. We inform them of what data is being used and why and only work with RBI regulated lenders who get access to the scores that help them make credit decisions, he added.
Moreover, the systems and processes of reputed fintechs are constantly audited by external auditors to keep the security top notch.
Download Money9 App for the latest updates on Personal Finance.