Article by Sandeep Anandampillai, CO-Founder & CPO, Crediwatch
As India’s business ecosystem continues to evolve, rapid changes in technology and finance have opened up previously unknown ways of making financing accessible to many more SMEs. Today small businesses can get access to cheaper formal credit and reduce their finance costs by opting into next-generation banking and business tools.
Approximately 50 million medium, small and micro enterprises (MSMEs) in India make up for USD 2 trillion worth of business every year., which is 29% of the GDP. Many small and micro enterprises in India do not have an optimal digital footprint. They often also do not have formalized business practices like their larger market counterparts including large private and public enterprises. So, a dearth of data results in higher interest rates for them in accessing credit, leading to a USD 1 trillion debt financing gap. This debt financing gap has a far-reaching impact not only on such MSME but also on our broader growth. Allow me to explain with a brief example.
USD 1 Trillion Gap: How could it impact the grand scheme of things?
Suppose you are an MSME business that needs capital to source raw materials, say, to manufacture face masks. Financial institutions such as banks and NBFCs simultaneously have the capital to provide you. However, since you are an MSME with low actionable data, they can’t assess you and hence, extend the loan. Here, two developments are occurring simultaneously. While you have not been able to source cash to meet an unmet market demand (hence, lost business productivity and subsequent growth), the bank has lost interest that its capital would’ve generated. With the overall debt financing gap of USD 1 trillion, which is about one-third of our economy, we can only imagine the possibilities of growth and expansion that this credit gap might be failing to meet. All of it is due to unavailability of relevant data points.
Here, a multidimensional credit scoring approach can empower the entire ecosystem to build digital trust and thereby, unlock superior efficiencies with formal credit. So far, old school credit underwriting systems have dwelled on asset pricing that relies on core data points viz. time in business, the scope of the industry, personal credit score, and annual revenues. Although the indicators play an important role but cannot solely determine the creditworthiness of a business, especially when the business environment is shrouded in COVID-induced uncertainties. Moreover, many lenders lack the requisite resources to continually monitor SMEs closely when it comes to credit monitoring.
Next-generation business concepts like a dynamic trust score have started to drive massive change and opportunity in the lending ecosystem. A dynamic trust score contains thousands of traditional and alternative data points and monitors them in near real-time. It indicates the overall health of the concerned entity by also assessing the creditworthiness of partners, vendors, and suppliers in the supply chain. Therefore, this dynamic trust score not only safeguards a lender’s interest but also enables businesses to better understand and predict their own financial health in relation to their peers. And of course, it empowers a thin-file business to access formal credit with more confidence.
Besides, new-age fintech solutions allow businesses to self-report timely payments to their creditors to help build and boost their score. Lenders leverage machine learning tools to assess the credit risk levels a potential borrower presents to them. A high-quality trust score enhances a business’s chances of getting loans at a less expensive rate. This is one of the key reasons why businesses should keep a tab on their trust score as they will be scrutinized for discrepancies and frauds against collateral. Business trust scores can influence multiple things about the financial support a company can get, including the value of funding, repayment terms, and interest rates.
Now, let us delve into the components of this new-age credit scoring.
An AI/ML-enabled multidimensional credit scoring uses a combination of standalone image analysis, consent-based data, public data, and peer comparison to better underwrite and price credit to this underserved segment.
How does Image Analysis work?
Image analytics for credit scoring leverages advances in image processing techniques and machine learning to derive insights on MSMEs for whom traditional analysis is not sufficient. Algorithms provide insights on entities by evaluating images of the entity’s physical establishment and this helps in trust scoring.
With the wide availability of mechanisms to capture images, such as smartphone cameras and high-resolution satellite imagery, the idea is to collate and correlate such pictorial insights with traditional financial and non-financial data points. When such correlation and learning come together, it helps in predicting the financial strength of other such businesses as well – even though they may be thin-file or lacking traditional data points for credit decisions.
Framework for a consent-based Analysis
Taking consent from the borrower, a lender can now extract valuable information and perform 360-degree checks at the time of underwriting across GSTN filings, ITR filings, and bank statements. Advanced credit intelligence platforms also perform automated checks. These analyze each of the documents and verify information across all factors to highlight discrepancies and correlate with the information obtained from images. Much of this is done in an automated fashion, using cutting edge technology, giving confidence to entities who are providing consent data because the data ingestion process is anonymised. This protects the privacy of entities while simultaneously becoming a game-changer for quick decision making in the banking industry.
Approach to Public Data Analysis
Advanced trust scoring may involve obtaining data from various public sources such as Regulatory Registrations, Sanctions Screening, Statutory Payments, Trade Information, Litigation checks, Media Monitoring, and Sentiment Scoring. The latest business platforms are capable of transforming and stitching the data from disparate sources to a single business entity using a proprietary singularity model.
Such Trust Scoring models have a significant impact on borrower segmentation. They help financial institutions segment the borrowers based on risk appetite. AI-based trust scoring models provide a composite risk rating score (Consent Data Score, External Data Score, and Image Analysis) for each potential borrower. This possibly makes it superior to the traditional credit score obtained from a ratings agency.
Using a combination of the above insights and trust scoring, financial institutions can efficiently evaluate the risk involved with each potential borrower and decide on lending – even if the entities might not have a traditional credit rating or credit score.
In the macroeconomic landscape, what this paradigm-changing development does is simple yet effective. It paves the way for channelization of credit throughout the market, especially across the bulging bottom of it. The end result is a higher growth trajectory wherein unmet demands are ably met by MSMEs, leading to an enhanced market performance led by the power of data!