What does it take to disburse a Rs 20 crore loan digitally?
A lot, but hear me out
Last week was revelatory.
We were presenting partners at the India Credit Risk Management Summit (ICRM) in Mumbai last week, which meant meeting new people and getting scattered but pointed views into the industry, its challenges and opportunities.
During the summit, a panel member mentioned that their goal is to digitally transform their lending organisation to the extent that they can digitally disburse a Rs 20 crore loan.
My first instinct was to dismiss it and respond with ‘not in this lifetime’. I didn’t, though; I nodded and made an expression that hopefully conveyed my disbelief and a deep appreciation for his ambition.
And honestly, that's the only statement that I’ve been ruminating over for nearly a week now.
In typical Big Picture fashion, I got back home and started doing some digging/wishful thinking to determine the chances of this panel member’s ambitions materialising.
I thought I’d jot down my two cents on this issue.
The current state of digital lending
The year 2022 saw booming credit demand, increasing adoption of all things digital, and the year the Reserve Bank of India (RBI) legitimised digital lending (in one way or the other, anyway). The impending Data Protection Bill, the Account Aggregator framework and the Working Group on Digital Lending (WGDL) to provide recommendations on how to run a digital credit ecosystem are all silver linings. But they’re also guardrails for the industry, so growth is organised.
Digital guardrails are great, but sometimes, they stymie innovation.
The RBI's guidelines place restrictions on current loan disbursement and repayment fund-flows, forbid granting credit on e-wallets, regulate how fees are collected by lending apps, require mandatory credit bureau reporting of all digital loans, regulate how customer data is collected and used by fintech companies, and limit first loss default guarantee (FLDG) arrangements.
Moreover, these rules primarily apply to sachet loans to promote wider inclusion. According to research provided by the industry's Fintech Association for Consumer Empowerment (FACE) and credit agency Equifax, personal loans remained the most popular product, accounting for 96% of the volume. Most personal loans were given for less than Rs 5000; however, in the consumer market, the category of loans with ticket sizes between Rs 10,000 and Rs 50,000 has had the greatest growth.
The promised land - disbursing 20 crores digitally
‘Not in this lifetime’
My instinct to say that came from the multiple regulatory guardrails that have been set up. The RBI’s digital lending guidelines are for smaller loans that are unsecured, are fairly easier to process and don’t require a relationship with your customer longer than a year at best. And yet, these guardrails are impenetrable.
Now, imagine the digital guardrails for a 20 crore loan that requires collateral, myriad proofs of existence, affordability and non-delinquency, double checking all of this, relationships with your customer that spans decades, the resources needed, risks to factor in, servicing and most importantly, trust. Phew.
Not in this lifetime, maybe.
But also, I have some calculated predictions that may work out in this lifetime. What are the factors that make a loan great? Relevant data, credit models, factoring in risk, regulation, pricing, NPA planning, monitoring and servicing post-disbursal and trust. The devil is always in the details, so let’s get into the nitty gritty -
The PSB-Loans-in-59-Minutes programme, introduced by the Government of India (GoI) in 2018, allows MSMEs to obtain collateral-free loans up to INR 5 crores. To that effect, the GoI has collaborated with over 21 public and private lenders.
While borrowers without collateral security can apply for business loans through the portal under the Credit Guarantee Fund Trust for Micro and Small Enterprises (CGTMSE) scheme, lenders can require collateral before disbursing loans.
However, the amount distributed through the programme only comprised 4.63 per cent of the overall bank credit for MSMEs in April 2022, worth Rs 17.63 lakh crores. In FY22, only 9,868 loans involving Rs 5,197 crore were sanctioned, of which 8,155 loans amounting to Rs 3,729 crore were disbursed.
There are two reasons why the GoI’s scheme hasn’t panned out as planned.
Lenders don’t have the technology for the straight-through processing (STP) of large business loans.
Data is scattered, and if it needs to be manually uploaded, document by document, it's prone to human error, therefore rendering these loans approved only in principle.
We’ll discuss part 1 later, but how can the data part be fixed?
A Public Credit Registry. The RBI’s mammoth project, still underway, intends to be the full database of data on all credit relationships in the nation, including payments, restructuring, defaults, and resolutions, from the point of credit creation to its termination. It’ll even include updates if a contract’s terms are changed during an ongoing credit transaction. This would make the registry an omniscient source of financial truth.
Integration of the TReDS platform with GSTN e-invoicing is a significant roadblock. Integration with GST will ensure the automatic uploading of all GST invoices approved by buyers onto TReDS. Subject to consent from MSMEs, this could create an environment of greater transparency and data accuracy.
Digitised land records. And not just digitising agricultural land that’s already underway as part of the Agristack, but a single, nation-wide window to determine property ownership. Currently, every state has its own portal where new properties are registered - an integrated approach will enable straight-through-processing for lenders in verifying whether a business property is rented or owned (or disputed).
All of these data sources, along with regular reporting to and integration with the Central Repository of Information on Large Credits (CRILC), can make the STP dream come true.
2. Straight-through-processing of commercial loans
Underwriters' manual work is cut by up to 80% when the straight-through loan processing methodology is applied to mortgage loan origination. Also, the loan processing time is shortened because human intervention is frequently avoided. Loan application data is put through the STP process, and a human review is only necessary if something in the application falls outside of the predetermined bounds.
And commercial loans, or loans above five crores (in this case, for MSMEs specifically), fall outside the predetermined bounds. Large business loans are complex - regulations, the correlation of participants in the lending life cycle, the disparate operational environment of products and complex servicing aggravate the complexity.
I will try to make a case for STP in large business loans to manage the loan life cycle and mitigate potential risks.
Bankers have been talking about STP for a very long time. Adoption is slow because larger loans bear more risk, and the basic human tendency is to verify and verify again. So let’s start with risk management -
In the wake of the Dodd-Frank Act, Basel III norms, Foreign Account Tax Compliance Act (FATCA) and other RBI-mandated regulations, banks must “aggregate, analyse, estimate and share data about customer transactions”. A risk model must account for risk elements such as industry, legal structure, macroeconomics, dispute history, etc. Remember that effective straight-through processing systems use Big Data from every industry, consumer demographic, and resource, including e-commerce data and regional economics.
The STP system can then access the shared database and gather the risk data required for an informed lending decision. Furthermore, as the lender's STP programme evaluates new applications, collects new data, and then shares the data, the information improves the integrity and accountability of all financial institutions. Open banking frameworks, like the account aggregator, will power STP, causing more and more lenders to choose STP for loan decisions as they gain greater trust in the data.
The data required to process a large business loan is closely tied to regulations. Here’s a brief look at all the data mandated to be processed for a large business loan and how regulation could possibly be tweaked to fit innovation in technology -
KYC - Digital lending guidelines allow for Video KYC, and lenders love it. It allows them to save resources on in-person verification. Currently, lenders are required to assign a resource to check for the liveliness of the video and ask customers to hold up their Aadhaar/Pan card properly so it can be read. A little bit of wiggle room in regulation can open up a whole new world of Video KYC, where all of these processes can be automated.
Banking - Unlike retail borrowers with one or two savings accounts, corporate borrowers have current and limit accounts. The celebrated Account Aggregator (AA) framework allows lenders to pull data from a retail borrower’s savings accounts. Currently, corporate borrowers need to manually upload bank statements for their current and limit accounts. Streamlining this process would mean bringing in current and limit accounts under the AA framework.
GST - GST data cannot be fudged. With the inclusion of GST data into the AA framework, lenders can now access a borrower’s GST data with a single pull.
Financials - These documents are manually uploaded, whether it’s audit reports or annual reports. A central repository which includes Income Tax Returns (ITR) and financials, will enable smoother STP for lenders.
Original, Seen, Verified (OSV) - This is possibly the most cumbersome process. Lenders are mandated to physically verify every document and premise of the business borrower. Not only does this delay the loan processing, but it also adds a possibility of fraud since most of this work is outsourced to contractors. How can lenders verify and validate information before they push money out the door without getting their boots on the ground?
There are some lessons from the west here - Truepic, a San Diego-based technology startup, employs controlled capture technology to capture a photo or video as soon as it is created. This enables users to authenticate an image's pixel data and certify the time, date, and location at the time of capture. Each image includes a URL, allowing anyone who captures the content to prove its authenticity in the future.
The technology enables banks to run images through 22 fraud detection and prevention tests within five seconds of capture. It also distinguishes between original photos and screenshots, which is an especially important feature for preventing applicants from submitting stock photos in place of their own.
It’s not as straightforward as it sounds - collateralized loans are complicated. The lender and borrower need to go to a registrar and submit a trail of ownership documents for all the properties they lease/own; once this is verified and stamped, there’s an additional process of lien marking that happens (basically, if that property needs to be sold/given up, the borrower in this case, needs an NOC from the lender) and then an Encumbrance Certificate (EC) that’s produced that is one of the final straws in sanctioning a loan.
It’s a lot, I know. Anytime there’s money involved, the trust deficit grows, and there’s room for multiple layers of verification and authentication.
Not in this lifetime
I set out to visualise what a world with awesome digital public infrastructure, technological innovation, and supportive regulations look like for a fully digital 20 crore loan. But I know that this may not happen anytime soon, but it’s worth laying out the landscape and working day in and day out to make this one step closer to reality. The journey itself is going to be worth it.
Hope is a wild thing. As much as you may wish to tame a wild thing, it will always resist and will do anything, at any cost, to break free. And my hope, much like the panel member, is to see a Rs 20 crore loan being disbursed fully digitally.
I’d love to hear your thoughts.
See you next week!