To Lend or Not To Lend? – Alternate Credit Scoring

Written by - Vibhav Bhat and Yash Prashant Bhave

PGDM Finance- TAPMI, Manipal

Long ago, lending activities happened by putting up on exchange different things, like cattle, houses, gold, etc. as collateral. We have seen movies with ‘Sahukars’ acting evil in this sector by seizing the collateral by offering unreasonable rates. However, with the evolution of formal banking one would expect people will approach banks for safe borrowing. But the ground reality of lending is different.

Lending Scenario Today

It is surprising to know that 94% of individuals in India do not have a credit score. More than 50% of SMEs don’t have credit scores either. Only 3% of SMEs have exposure to institutional lenders. The golden rule of lending is to check borrowers ability and willingness to repay. However, in case individuals or entities don’t have a credit score assessing this becomes difficult and thus lending is riskier.

To run a business either a Kirana store or an SME manufacturing entity, one requires funds. Since these entities find it difficult to obtain it from banks they usually end up borrowing from “Sahukars’ “, who charge at unreasonable rates. To resolve this issue and to maintain parity in interest rates offered, the government is focusing on financial inclusion. WEF’s report highlights that increasing 1% financial inclusion can lead to an increase in GDP by 0.03% a year. This is another reason the government is pushing the fintech entities who are there in the lending space. However, to lend to the underserved segment it is important to find out the creditworthiness of the borrower. This can be done using Alternate credit scoring method (ACS).

---

What is ACS?

ACS is a way of scoring borrowers with thin or no file data in the formal economy. It uses various data sources to analyze the individual or entity.

FinTechs are technologically robust. They use various data analytics tools to analyze unconventional data. By using the outcomes of the analysis banks or NBFCs can decide to lend or not for a borrower.

There are P2P (Peer to Peer) platforms that facilitate the lending activity by connecting the lender with the borrower. Here, ACS plays a major role to substantiate the creditworthiness of the borrower in front of the lender. Along with wide availability of unconventional data, it brings the disadvantage that is giving false positives which makes it riskier. It is the responsibility of the analyst to check for such scenarios and highlight them thus making algorithms better and lending safer.

---

What data does ACS leverage?

To evaluate the credit worthiness of borrowers who cannot be assessed on traditional scoring alternate datasets are used. The datasets that can be leveraged differ across geographies and depend on the behavioral patterns. The following photo depicts a few of these datasets that are currently being utilized.

Figure- Alternate Datapoints for credit scoring

Source- World Economic Forum Article

---

How is alternate data processed?

Given the amount of data accessed on these alternate datasets are humongous, sifting through this manually is time consuming and bound to produce incorrect results. Machine learning algorithms on other hand are leveraged to run through them and come up with useful insights. Machine learning not only reduces the turnaround time but also uncovers useful information on borrowers which can be used to rate them on intent and ability of repaying. This is useful to uncover the risk of lending.

ACS is promising especially in emerging markets where much of the population has no traditional scores but it cannot replace financial-based scoring. ACS employed alongside traditional scoring has much more potential to assess the lending scenarios. However, capabilities have to be merged on the side of banks to incorporate this and carry extra risk in their portfolio.

---

ACS Application and benefits

Currently ACS based scoring are used by multiple P2P lending platforms in India. Given a large portion of borrowers that require credit do not get it through banks because of lack of traditional scoring capability and flexible lending requirements, ACS is used by multiple FinTechs to increase financial inclusion. RupeeCircle, a P2P lending platform, has developed a proprietary credit scoring algorithm that leverages 5000+ data points to create a borrower profile for a thin file customer. It incorporates multiple datasets such as credit history, financial ability, social behavioral patterns and others to assess the intent and ability of the borrower. Based on this analysis a borrower is categorized on a risk scale from A-F (F weakest). On the rating the interest rates are decided and disclosures of the risk profiles are made to the lenders to maintain transparency.

This creates a conducive relationship on both sides. Lender knows the profile of the borrower and the risk associated. For a borrower looking for credit it allows him to borrow at reasonable rates and meet his/her goals improving the financial inclusion in the country.

---