Specializing in new ventures, AI, deep technologies, AML, risk intelligence and more, Meghna is a serial entrepreneur and has founded an AI company that uses deep learning tools to generate insights about private businesses.
It takes an excessive amount of paperwork for Micro, Small, and Medium Enterprises (MSMEs) to justify their financial standing to avail of a loan, mortgage, or other banking services. Credit underwriting for businesses relies heavily on their time spent in the market, industry potential, personal credit scores, and annual revenues. A dearth of data around the same results in no credit or a higher rate of interest. For banks, extending the credit means earning interest and thus, superior profitability.
Approximately 50 million medium, small and micro enterprises (MSMEs) in India make up $2 trillion worth of business every year, which is 29 percent of GDP. A majority of these enterprises have a minimal digital footprint to get their creditworthiness assessed. This has led to a $1 trillion debt financing gap. This debt financing gap has a far-reaching impact on broader economic growth. This is a sizable issue for a developing country like India which is currently reeling under the economic pressure of the COVID-19 outbreak.
Here alt data and technology can open up new avenues for SMEs and financial institutions.
What is Alternative Data?
Alternative data is additional information that is extracted from sources other than the traditional data sources. It can include a borrower’s payment history, such as cell telecom and utility bills, GST filings, rentals (if any), cash flow details against personal bank accounts, a customer’s use and repayment of certain alternative loans, and so forth. This data may look unrelated but gives in-depth insights into their attitude towards financial commitments and additional economic loads.
For banks and NBFCs, alternative data is increasingly becoming instrumental in serving thin file businesses that may otherwise not be approved for loans. And for enterprises, meeting expectations on these parameters promotes a healthy credit score and a fine digital footing. This information is further used in screening out fraudulent practices and better servicing the customer as well as loan repayments. Moreover, it can be applied for account monitoring purposes. It specifically becomes crucial amidst the pandemic-induced uncertainties in the current business landscape, as no financial institution wants to fall for bad loans and NPAs in its loan book.
Pivoting around a vast amount of personal data, collecting alternative data requires the approval of the borrowing entity. Alternative data is not available in the public domain, which needs the customer’s approval that his/her data be submitted to the third-party for scrutiny. Hence, a consent-based framework becomes crucial as it also ensures data privacy at all stages. Here’s how it works:
- Get approval from the applicant to analyse its personal data for a better digital footprint.
- Based on the credentials shared, the system fetches the data from the service provider.
- These details can be checked for relevant information such as frequency of payment defaults, due amount, the amount paid, and others. For example, data points from the telecom operator can be analyzed to understand the user’s credit profile. On-time payments from the same payment mode against a heavy call volume are fair indicators of the stability of the profile. On the other hand, inconsistencies in running regular and basic expenses can act as a red flag. Use cases of telecom data include analysing default probability, operations of the company, network size, telecom bill efficiency, chances of using black money, and even knowing the tech-savvy quotient of the borrower.
Overall, default probability, operations of the company, network size, telecom bill efficiency, transactions involving illegitimate revenue sources are some of the aspects that a lender can analyze with alternative data. And, technology is an integral part of it.
Alternative data and New-Age Technology at Play
The state-of-the-art technologies such as Artificial Intelligence (AI) and Machine Learning can extract relevant information to provide actionable insights from high-volume unstructured datasets. Feeding and processing this data through the systems of fintech platforms can render extended insights into a borrower’s overall financial health in real-time. Machine Learning algorithms perform data analytics to identify signs of distress on these borrowing entities or patterns that relate specifically to credit risk.
AI-backed alt data analytics can track thousands of data points in real-time to develop comprehensive credit reports. These scalable systems require no human intervention, thereby increasing the likelihood that the decisions made will be free from cognitive biases. It could indeed pave new realities of credit scoring in the future and avert the issues plaguing the banking system birthed by the non-availability of data.
