[ad_1]
The Indian banking system’s bad-debt level is the lowest in over a decade, but the sector’s regulator, the Reserve Bank of India (RBI), is ever vigilant. It has spotted fault-lines in unsecured retail lending. Specifically, small ticket personal loans (STPL) with ticket sizes below ₹50,000 are showing at least 2-to-3 times the delinquency levels of other retail loans. To mitigate this risk, RBI has so far focused only on lenders, which are now expected to keep higher equity levels for all unsecured consumer loans. However, RBI may have missed out on crucial stakeholders in this exercise: STPL borrowers. It is doubtful whether such loan-takers, many of whom have benefitted from financial inclusion only recently, are fully aware of the limits of their debt-servicing ability. The deluge of STPL loans made available by a subset of lenders may lead many of them to over-rate their ability. Besides, they may not be aware of the consequences of a damaged credit profile. It is here that RBI could step in with an awareness campaign. After all, it is the loose lending practices of some players that have raised the risk, and not the country’s financial inclusion drive.
A stitch in time: RBI has run awareness campaigns for customers on issues ranging from grievance redressal to fraud prevention. The time is ripe to sensitize borrowers to the problem of over-indebtedness and the importance of a good credit profile. Many borrowers are swamped with loan offers, even for impulse purchases, and some of them may end up taking on more debt than they can handle. While an over-abundance of short-term credit supply may allow them to roll over their loans, at least some are only a single shock away from default. If we see a widespread surge in retail pay-back failures, damaged credit profiles could undo the achievements of financial inclusion.
Credit access is a challenge for borrowers with poor credit scores. Plus, there may be emerging socio-economic implications. Increasingly, employers are requesting prospective employees to share their credit reports. In some cases, these are sought even for matrimonial alliances.
Financial inclusion is not the issue: A narrative has emerged that banks and large non-bank financial institutions do not easily lend to loan applicants who do not have a credit-bureau record. Such individuals are called new-to-credit (NTC). This narrative is an exaggeration, though. Established lenders do lend to NTC-clients. Loan decisions in their case requires many more human touch-points apart from financial information to assess their debt-servicing ability and willingness. This could take days instead of hours, as is usually the case with loan applicants who have a long bureau record.
As such, NTC borrowers do not exhibit higher delinquency rates than an average borrower with a credit history. Yet, credit risks would emerge if over-enthusiastic lenders jump forth to underwrite NTC borrowers within minutes without suitable techno-analytical capabilities and guard-rails.
Alternate data is filling gaps: Most established lenders have been able to maintain a balance on the risk-growth tightrope of retail lending, thanks to alternate data. With the consent of customers, lenders can access plenty of data on them for loan approvals. Account aggregators are playing a role in gathering information such as bank statements, while telecom usage bills also come in handy and other indicators of behaviour can be captured via mobile phone apps.
A few lenders are found to be doing a good job of leveraging alternate data for deciding on NTC loans. Such players have access to machine learning (ML) models that are explainable and stable, although model-based lending decisions need to be reviewed diligently and the models kept updated.
Why should RBI step in? The competence shown by a few lenders cannot be generalized across the industry. Several lenders are found to be following worrisome practices. Among their minor problems is the deployment of black-box models that use hundreds and sometimes thousands of variables to assess credit risk, with little understanding of how most variables impact the measurement. In a bid to put in exotic variables, some are using metrics that border on biased opinions as opposed to empirical facts. The presence of specific social-media apps on the mobile phones of borrowers, for example, has been known to impact risk scores. Models also exist that take into account the morning-alarm time set on people’s phones. Lack of statistical rigour leads to counter-productive risk models. Conventional risk factors include bounced cheques issued by applicants, low average bank balances and credit-bureau flags raised for borrowers who have taken out too many loans in the past 3-6 months; these are used by saner models.
Next is the relentless hard-sell strategy adopted by some players. Easy availability of credit, often driven by badly designed risk models, allows weak borrowers to pay off one loan with money received from another. Sometimes, it’s the same lender that ends up giving top-up loans. However, loan ever-greening is often a sign of borrowers having taken unsustainably heavy debt burdens.
Credit deterioration is a direct outcome of such questionable risk practices. Thus, RBI must educate borrowers on how best to estimate their debt-servicing ability and not over-borrow. No matter how strongly RBI pushes lenders to limit lending, it cannot improve the behaviour of all lenders.
[ad_2]
Source link