Here Are The Best Ways For Lending Companies To Maintain Low NPAs
As bad loans continue to plague the country's economy, it is now critical for lending institutions, to be careful about allotting loans and stringent about having a robust collection mechanism.
The gross non-performing assets (NPAs) of Indian banks was calculated to be Rs 8,40,958 crore last December. These bad loans as on December 31, 2017 due to loans to industry were at Rs 6,09,222 crore, dues from services sector was Rs 1,10,520 crore; Rs 69,600 crore was from agriculture and related sectors, Rs 14,986 and Rs 36,630 crore from other non-food credit and retail loans respectively.
SBINSE, India's chief lender, accounted for the maximum amount of gross NPAs - Rs 2,01,560 crore. Others included, Punjab National Bank was at Rs 55,200 crore; IDBI Bank - Rs 44,542 crore; Bank of India - Rs 43,474 crore; Bank of Baroda - Rs 41,649 crore; Union Bank of India - Rs 38,047 crore; Canara Bank NSE -2.56 % - Rs 37,794 crore and ICICI Bank - Rs 33,849 crore.
With this grim reality plaguing the country's economy, it is now critical for lending institutions, big or small, to be careful about allotting loans and stringent about having a robust collection mechanism.
Important To Gauge Capacity & Intention of Customers
When you lend and create a portfolio, the portfolio has an inherent risk. The hallmark of a good portfolio is one where risk and reward is commensurate and NPAs are kept at an optimal level. Traditionally the way to assess risk for a given customer was to look at their credit history and give loans on this basis. However, this approach has a drawback. It's biased against first time borrowers and it doesn't evaluate an existing borrower on his ability to repay in the future.
Once a customer defaults on a payment, NPAs can be avoided by an effective mechanism that centres on tele-calling, dunning and door knocks. Entrepreneur India got in touch with players in the lending space to find out ways to maintain low NPAs in the company balance sheet.
It's Pure Science & Algorithm
According to Monish Anand, CEO and Founder, Shubh Loans, "Quite often, the perception of collection is wrong. People think it's about muscle power when, in fact, collection is pure science. It requires a differentiated strategy for early and chronic defaulters."
"The core aim of Risk Management is to assess and triangulate both capacity and intention of prospective customers. In-case of existing borrowers, the intention can be largely assessed through customer's bureau score and his/her past performance. However, for new borrowers data science is opening up doors by identifying proxy factors that indicate a customer's intent and ability to repay a loan. Turns out mobile phones are a great source of proxy data. From the way a customer fills the online loan application to the number of contacts she has in the phone, including their bank transaction history via SMS - all come together to build a well defined risk profile," adds Anand.
In Shubh Loans experience, over 90% customers make repayments in early stage of defaults by just effective telecalling. At a later stage, customers are counselled face-to-face where they are made aware of the need to be non-delinquent on the account.
Satyam Kumar, Co-Founder and CEO, LoanTap, follows leveraging proprietary algorithms to generate customer credit score. "Loan applicant shares his income and current obligations on our website and we convey values of loan amount, interest rate and loan tenure at the same time. If customer finds our provisional offer suitable, then he/she shares required documents with us. These details are then analyzed by algorithms to effectively judge the customer's creditworthiness. Such robust credit analysis procedure helps us in keeping NPAs at check," elaborated Kumar.
Manish Khera, Founder, Happy Loans, believes there are only two ways to run a lending business - Coercive Collection or Prudential Lending. He espouses the latter one.
"We assess over 1000 variables about the merchant to underwrite a micro business. We are learning about the customer and his unique behaviour every day. Thanks to the technology innovations like artificial intelligence and machine learning, our learning process is now more practical, efficient and fast. I believe, process efficiency is the key for digital lending - and have automated the repayment from the wallet or POS sales that mitigate the risk. The merchant is not required to do anything to repay, the entire process is digital and automated," elucidated Khera.
Get Right Customers And A Robust Collection Process
To ensure that our NPAs stay low, Bhavin Patel, Founder & CEO, LenDenClub has developed an all-inclusive self-learning credit and collection model. For reducing borrower defaults Patel prescribes as two-way mechanism - choose right borrowers and set up robust collection processes. To choose right borrowers, he has made sure that his company comes up with timely analysis of the borrowers' behaviour and reflect relevant changes. His credit model includes traditional parameters like credit score, stability, education etc. along with non-traditional parameters like social media verification, online buying patterns etc.
"We have also come up with some unique parameters connected to borrowers' family composition. We make sure that every borrower we list on our platform goes through thorough fraud checks clubbed with mandatory physical verification of the borrower's residence. We have implemented machine learning within our credit model which automatically helps us identify and weed out borrower profiles which are similar to the ones that have defaulted with us in the past," he disclosed. In last 2.5 years, LenDenClub's credit model has gone through around 20 revisions to reach current gross NPAs of 2.96%.
The second important part, as stated by Patel, is to set-up correct collection mechanism. In lending business, borrower defaults at times but pays back as well. If the lending company doesn't reach out to the borrower within stipulated timelines, there is a high possibility of default. At LenDenClub, the tele-callers are trained on the language, tone, and communications so that they convey right message in a right way.
"From our initial days, we have set standardized process of recovery where our team knows when to shift a case to external recovery agency or legal agency for further proceedings. Our collection process is quite streamlined and we follow the same across geographies," concluded Patel.