Business credit in India is biding its time Moneyball
The big booms in corporate lending are not limited to state banks. They occur regularly in private banks, mutual funds, and private equity sponsored non-bank finance companies. It is questioned whether there are chronic systemic deficiencies in the risk processes of business loans. Only a fraction of these defaults can be attributed to borrower fraud and corrupt bankers. However, an undue generalization of this narrative may have prevented an introspection of existing risk practices. Gradual changes have taken place, such as increased use of data, but only a few large banks have fundamentally redesigned the lending process. For many banks, the business lending process remains highly subjective, sometimes guided by unverified rules of thumb and selective readings of data. Even the tailor-made “expert models” that are sometimes used have limited risk prediction capacity.
Moneyball in business loans: The Billy Beanes of business loans must come forward. Billy was the legendary manager of the Oakland Athletics baseball team. His exploits in using analytics to select players for his squad and his subsequent success are the subject of the book, Moneyball: The Art of Winning an Unfair Game, and a film of the same name. Baseball teams have a budget to buy players for the season, like our Indian Premier League. Before Billy introduced player analysis, selection and pricing was determined by baseball experts with decades of talent scouting experience.
When choosing players, besides basic performance stats and their immediate success, the personality traits of the players as perceived by these experts had an impact on their selection. Parallels can be drawn between the way underwriters look at financial statements and the perceived quality of management in selecting customers on credit. In the film, a talent watcher who dislikes the analytical approach says, “Baseball is not about numbers, it is not science” and “There are intangibles that only baseball people understand ”. Replacing baseball with business loans may reflect the thoughts of a hypothetical underwriter with doubts about analytical intervention in high-priced loans. The film dramatized certain facts. The data analyst had a degree in economics but little prior knowledge of baseball. In reality, analyst Paul DePodesta was a baseball scout himself. Ultimately, sports analysis improved the decision-making skills of talent watchers like Billy Beane, but did not replace them.
The human element versus expertise: It may be interesting to note that predictive models alone are unlikely to beat the top underwriter working at the height of attention span, without pressure to meet quarterly credit goals. But consistency issues can arise. Research suggests that in marginal cases, with no black-and-white response, the underwriter’s mood matters. On high mood days more loans are approved and higher defaults are observed for these loans. (In The Mood For a Loan: The Causal Effect of Sentiment on Credit Origination, Agarwal et al, 2012)
The challenges of “ expert models ”: Some lenders are aware of the problem of inconsistent decisions, but continue to rely heavily on qualitative aspects. Sometimes this leads to the use of quasi-quantitative assessment techniques, such as expert judgment models. These models use hard information (financial statements) and soft information (management quality). The soft aspects, being qualitative, are difficult to verify and their interpretation depends on the individual. Research suggests they can lead to worse loan decisions, especially if the person collecting them is pressed for time. One study found behavioral nuggets: For example, when loan officers have earlier sales experience, the confidential information collected tends to be interpreted with a more optimistic bias. (Making sense of soft information: interpretation bias and loan quality, Campbell et al, 2018)
As such, hybrid models that use qualitative and quantitative variables may not always be statistically robust. Qualitative data often does not stand up to statistical examination of its predictive power. If one adds in soft information collected in an inconsistent way, it is not surprising that it has less predictive power than models using hard information.
Start by measuring the quality of decisions: First, you may like to identify and measure the shortcomings of the traditional underwriting approach and determine what works and what doesn’t. The final “go” versus “no go” underwriting decision is preceded by several elements of prediction and judgment. Some of these items are industry outlook, financial statement projections, future debt requirements and the quality of management. The values of these components continue to evolve as the credit call moves from the underwriter’s initial recommendation to a credit committee, and then higher. While most lenders are able to track the quality of the final decision by loan default rate, some are unable to assess the quality at every stage. For example, if the financial projections are very far from the actual results, even though the account is not overdue, this suggests a weakness in the process.
A high failure rate can be the result of poor underwriting or poor economics. Likewise, a low default rate does not always mean high quality underwriting decisions. Knowing the gaps will help design an accurate analytical intervention. The goal, as in Moneyball, is not to replace the underwriter, but to make better decisions.
Deep Mukherjee is Visiting Professor at IIM Calcutta and Risk Management Consultant
Never miss a story! Stay connected and informed with Mint. Download our app now !!