Heavier risk weights can’t save lenders from every fire | Mint

Heavier risk weights can’t save lenders from every fire

For some new age lenders with significant UCL exposure and NPA rates above 5%, even 11.25% capital is not enough.
For some new age lenders with significant UCL exposure and NPA rates above 5%, even 11.25% capital is not enough.


  • Consumer lending has seen risks but bigger capital buffers, as RBI has asked for, may not be a solution to that problem.

India’s unsecured retail lending has some emerging worries. However, systemwide high delinquency is not among them. Non-performing-asset (NPA) rates by the 90+ days-past-due (DPD) rule range from 0.8% for personal loans (PLs) to 1.6% for credit card receivables, as shown in a recent TU Cibil report. These products account for 93% of unsecured consumer lending (UCL). Given the burgeoning asset growth, these coincidental NPA rates tend to understate riskiness. Loans given in the last quarter will not jump to NPA status at the quarter’s end. However, if one even considers the lagged NPA rate: i.e., current year NPAs as a proportion of exposure 12 months back, it is 2.9% (first quarter of 2023-24). It is an improvement from the 3.3% lagged NPA rate of 2018-19 (bit.ly/3NkJM32). When the Reserve Bank of India (RBI) reduced capital requirements for PLs in September 2019, the UCL delinquency numbers were not materially different from the latest ones. Yes, a spot of bother exists for small ticket personal loans (STPL) of under 50,000, but that is just 0.3% of the retail portfolio. This may be a symptom of a deeper malaise that an enhanced capital-cushion requirement, by itself, may not solve.

Too much of a good thing: Indian banking follows the standardized framework for capital estimation. The regulator mandates capital by loan type for retail exposure. Most large economies estimate capital by the Advanced Internal Rating Based (AIRB) approach. While it is not the panacea for banking ills, such regimes have the data-analytical ability to measure risk and capital requirements better. Estimation of Economic Capital (EC) using historical loss data is common, where EC is the amount of capital which can cover an extreme loss of one-in-1,000 odds (loss levels that can happen once in a thousand years, i.e.).

Previously for UCL, when its Risk-Weighted-Assets (RWA) norm was 100%, lenders kept 9% capital (excluding capital conservation buffer). So, for a 100 loan exposure, they had to keep 9 equity. Last month, RBI upped that weight to 125%, so lenders are now expected to keep 11.25 of capital for the same 100 asset. Yet, EC estimated for Indian bank’s UCL suggests that keeping 5% to 8% of exposure as capital is ample for most of them to handle a once-in-a-1,000-year loss. For major banks whose UCL delinquency rate barely hits 1%, RBI’s revised capital ask may thus appear too high. A data-backed explanation from RBI for its move would have been welcome.

Potential unintended consequences: For some new age lenders with significant UCL exposure and NPA rates above 5%, even 11.25% capital is not enough. Thus, the new RBI guidance might not be demanding sufficient capital of the weakest of lenders, while adding to the buffers of already well-capitalized banks. Further, this move may cause capital misallocation. Using historic loss data, if EC is estimated for a portfolio of commercial, MSME or agri-lending, one may see an EC requirement of 10% to 15% of nominal exposure. Because of mandated risk weights, typically banks end up keeping aggregate regulatory capital of 5% to 9% of exposure for such loans. As such, most banks do not estimate EC at the portfolio or even bank level. Given the higher capital ask for UCL, banks may push capital towards a higher-risk portfolio while reducing exposure to low-risk retail loans. One must track whether otherwise well-intentioned macroprudential interventions such as increased risk weights for UCL ends up increasing systemic risk.

Rethink needed on UCL lending: In the last 10 years, UCL has been driven by salaried borrowers employed by large companies (‘Category A/B’ employers). Underwriting frameworks and risk models on this are quite robust. In the last few years, the proportion of salaried borrowers working for smaller companies (‘Category C/D’ employers) has increased. Also, unsecured lending to self-employed borrowers has gone up. Despite pockets of risk excellence, gaps remain in the risk practices of such borrowers. Some new-age lenders may be using badly designed risk models to lend. These tend to misuse machine learning (ML) and alternate data. Perhaps future portfolio blowups will be attributed to ML models while the culprit may be the poor analytical maturity of such lenders. Delinquency within the STPL segment is possibly suppressed right now by liquidity infused by credit oversupply. Many over-leveraged borrowers may be just a shock or two away from going delinquent. Knee-jerk reactions by STPL-focused lenders may actually trigger a spike in delinquency through a sudden reduction in credit flow to such borrowers. Thankfully, the scale will be contained.

All UCL lenders need to reassess and reimagine their UCL underwriting skills. Their systemic capability must not fall short on grasping household level debt, the finances of gig workers, unemployment trends and the strength of links between employers (particularly category C/D) and employees. Risk models need to be tested for their ability to withstand a crisis. Model validation and maintenance need attention. Boards of banks and senior management must play a more active role to mitigate such emerging risks. Neither having an excess capital buffer nor a prudent regulator relieves them of their responsibility to keep risks in check.

Of course one must act before the house is on fire, but if fire is likely from loose electrical connections while one is preparing for a wood fire, the house may still remain at risk.

Catch all the Business News, Market News, Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.


Switch to the Mint app for fast and personalized news - Get App