Click Here for Sector Update
Credit risk from rules-based models show red flags
Over the past few years, lenders increasingly have turned to rules-based credit engines in a bid to bolster growth. On one hand, they have helped to achieve scale while being efficient while on the other hand, outcomes are yet to be tested for durability. The emergence of credit engines has been tested on limited historical data points, thereby needing further validation, especially in dynamic market conditions. In this context, there is a likely emergence of initial signs of vulnerability in these models. One derivative of such a model is colending, which we believe is inherently a win-win proposition, but the emerging trends do warrant caution.
While colending is win-win proposition, vulnerabilities exist
Colending has been a big area of growth with AUM likely to touch INR 1tn as per CRISIL Ratings. The underlying architecture of colending makes it a win-win for the ecosystem comprising banks, NBFC and fintech. On one hand, it allows NBFC to grow in a capital-efficient manner, enabling access to bank funding. On the other hand, it provides banks access to niche customers while sharing risks. The basic operating tenet of colending looks tenable. That said, the underlying principle of risks-sharing, model-based lending and pricing still have not been validated for wider datapoints. Moreover, certain practices, such as not following the 80:20 rule in spirit wherein seller eventually securitizes a 20% portfolio in its book may potentially attract regulatory attention sooner than later.
FLDG still is not a full-proof concept; tail risk on the anvil
There have been discussions on first loss default guarantee (FLDG), which faced regulatory interventions with the RBI allowing FLDG, up to 5%, to help the ecosystem. But we need to answer two queries: 1) Is FLDG a full-proof concept and is the risk being adequately priced in? and 2) Is FLDG being followed in letter & spirit, which is paramount for the regulator? We believe the answer is rather negative in both cases. First, FDLG has not been tested to adequately price the risks. CRISIL Ratings states more than 30% of colending has taken place in the unsecured personal loan segment wherein gaps are visible on credit cost. This portfolio currently faces a double whammy of slower growth and rising delinquency, which may crystallize in the near term, and feed into adverse experiences. Second, the emergence of risk premium (other means) to offset for FLDG may not be tenable.
Unlikely to see a snowball effect, but a few hiccups likely
Given the nature of the portfolio - largely retail & MSME as per CRISIL Ratings -- we see more localized effects than a contagion one. We believe a few business models might be at the liquidity risk, given that an adverse experience by one may lead other lenders (liquidity providers) to pull the plug. Also, diversification challenges (operational & technological) will entail that one event with one lender will make it difficult for smaller entities. We note at this juncture we are not too concerned on the MSME & secured portfolio, but expectations may get corrected as the model is adjusted. |
