How Unsecured Loans are Embracing Small Businesses? MSMEs only received 17.4 per cent of the total credit by formal lending entities in India but all this is about to change

Borrowing has never been as easy as it has become today in India. If you had to borrow during the yesteryears, the ordeal was quite real and identical, irrespective of the bank that you were borrowing from. First, you had to take a day off to go to the bank, submit your loan application, and furnish it with additional paperwork. Then, you had to constantly visit the bank to check the progress of your application, ultimately, to know that the application has been rejected post-assessment.

So, after losing around one to two months and taking multiple leaves from your office, you're back to square one. And this was the case for consumer loans. Business loans, specifically micro and small business loans, were a different ball game altogether. The government did its best through cooperative banks, SIDBI, NSIC, and various schemes like CGTMSE, but they didn't have an optimal result on the ground as was being anticipated – considering the fact that Indian MSMEs still face a credit gap of $650 billion.

Thankfully, India has been changing lately and so is the way how loans are processed within the country. We'll get into that, but first, let us have a look at the industry.

How Technology is Addressing a Long-Standing Issue of Credit Analysis?

According to the Economic Survey of India, whose findings were released last year, MSMEs only received 17.4 per cent of the total credit by formal lending entities in India. The remaining (82.6 per cent) was facilitated to larger enterprises. This is when MSMEs contribute about 32 per cent to the nation's GVA (Gross Value Added) and about 45 per cent to the overall exports and manufacturing output.

MSME credit penetration turns out to be quite challenging. It is because the MSME sector lacks viable data points to back a loan. A lender has to analyse multiple points before facilitating credit to any business entity. This includes the credit history of a business, its financial health, market vertical that it caters to, business association data and relationships with suppliers. Other factors considered are compliances such as enrolment with GSTN, ZED rating, certifications, loan defaults from the circle, and market insights pertinent to the business model just to name a few.

The relevant data is easily available and verifiable for larger businesses, but this is not the case for the smaller ones. What adds to the situation is that the formalization and digitization of small businesses in India is still an ongoing process. This makes the relevant data gathering considerably difficult, time-consuming, and yet a process prone to errors. Given this, there are about Rs. 80,000 crore worth of stressed loans within the MSME sector, and this figure is when only 10per cent of Indian MSMEs have access to formal credit.

So, lenders ideally avoid facilitating credit to the MSME sector or keep their relative share to bare minimum, so that they can avoid subsequent NPAs originating in their loan books. And this is precisely where Artificial Intelligence, or its subset "Machine Learning', is making a difference.

What is Machine Learning?

Machine Learning is the branch of Artificial Intelligence that leverages statistical techniques to help computer systems "learn' by themselves. The technology establishes statistical correlations and gathers empirical evidence to back its findings rather than basing it on pure logic or absolute parameters. Today, forward-thinking lenders are using the technology to automate loan assessment and make it more accurate. This helps in reducing the man-hours spent on such intricate processes as well as related errors, omissions, and negligence. At the same time, the technology helps in increasing the credit efficiency as well as the overall market productivity.

As one of its biggest merits, Machine Learning eliminates the redundancy from the credit analysis process that usually develops over a period of time. In general, many creditworthy applicants are neglected due to unrefined classification and predominant biases about an industry vertical or the loan application itself. Simultaneously, some of the unworthy applications are approved due to such classifications, adding to the overall NPAs. Since Machine Learning is a data-driven approach, it dynamically develops itself based on the historic as well as rapidly generating data for unmatched precision within loan analysis.

Battling NPAs: The Sheer Ingenuity of Tech-driven Loan Disbursement

When complemented with Big Data analytics, another technology that is able to analyse high-volume, rapidly-generating data to draw critical insights, Machine Learning can virtually make wonders happen. It can travel beyond the traditional limitations including the absence of data as experienced in the case of smaller businesses. This approach is being used to analyse the NPAs and the principal reasons for them to turn into defaults. A number of statistical datasets can be used and correlated including business performance, the market condition (of the relevant vertical, its associated verticals, and the broader market), historic projections, circle-based data, transactional data through GSTN, Aadhaar data through AEPS, profile of the business owner, and so on, to better gauge the reason behind the loan default and where the loan assessment lagged.

On the other hand, this approach is also being used on best-performing accounts to draw similar insights. Predictive analytics, at the same time, is helping understand the potential of individual verticals and other factors that affect the loan assessment process. With every successive default or high-performing account, the approach automatically gathers more data points and enhances itself.

Using this approach also eliminates human intervention from the analysis, and thereby errors, omissions, and negligence that usually arise given the intricacy of processes. Some MSME-based fintech platforms that leverage Machine Learning – despite their systems still being in an embryonic stage – have been able to keep their NPAs well below 3 to 5 per cent as compared to 20.41 per cent NPAs that scheduled commercial banks have in the gross advances.

This has been enabling small businesses to easily avail loan collateral-free loans without experiencing the red-tapism of yesteryears. We can, therefore, conclude that the technology is not only relevant for loan assessment of smaller businesses; it is gradually becoming indispensable. As the tech adoption increases in loan processes, thanks to the forward-looking digital lending platforms, it can be said that the existing credit gap will eventually disappear, and so will the NPAs from the MSME sector. All we have to do is wait till that happens.