Indian banking industry can weather stress in unsecured loans, says BCT CEO

Vaidhyanathan is an fintech entrepreneur, whose company developed risk management software modules for the banking industry to red flag issues that could affect the credit portfolio. In her interview with Deccan Chronicle, she touched upon various issues that the banking industry is confronting today.

Excerpts from the interview:

Q. Bad loans can happen in two scenarios — when the borrower is unwilling to pay or when he is unable to pay. So it looks like you are focusing more on the borrower’s willingness to pay.

Unwillingness to pay is fraud, while inability to pay is stress. We focus on both. If you're lending to an industry, let’s say agriculture, which is seasonal and if indicators suggest that there are chances of low rainfall, farmers’ inability to pay would be genuine as they are confronted with a crop problem. However, their inability to pay could be short-term, not long-term, in which case you could seek other additional collateral. If it's a term project, you could say you will not lend it for a particular period, but support it at a later stage. So there are multiple actions which banks look at taking to address these issues.

Q. So how do you rate the Indian bank sector’s credit portfolio?

RBI is a very innovative central bank regulator and I would say comparable to one of the best in the globe as far as regulation is concerned. So from a perspective of the credit reviews, I think Indian banks are quite strong, but the fraudsters are coming up with new methods of frauds.

NPA was the biggest issue 10 years ago. The regulator strengthened its oversight and now you don't hear about NPAs as much in the press now anymore. It's no longer as prevalent as it used to be. Now you are hearing a lot about mule accounts. There are 400 to 500 crore mule accounts in India. So the causes of fraud are different, but the fact remains that banks continue to bleed money. So what form fraud assumes is different. Therefore, the credit synchronisation is not static. It needs to be ongoing and dynamic.

Q. In most NPAs, barring those who took loans with fraudulent intent, stress comes first and fraud comes later.

Yeah, correct. So our focus is equally on the ability to pay. We use a lot of big data leading indicators to track commodity prices, input prices, output prices or any other issues. For example, during a particular season, the price of electrodes used for steel manufacturing went up by 100 times. That was a major red flag for the steel industry. In regulated sectors, you cannot say that I will increase selling price because input cost went up. So we use a lot of third-party indicators to predict stress.

Q. In India, there are indications of stress levels rising in unsecured loans.

An unsecured loan is always stressful business from a bank's perspective, because that is not backed by anything that you can go after. So that is always considered to be one of the highest stress points for any bank. Are we continuing to see default in the portfolio? It again depends on the borrower profile, loan type, and industry. However, there is a considerable amount of tightening now (for unsecured loans). Because when money is in short supply, it goes to the right people, but when money supply increases, it goes to everybody, in which case the credit review becomes poor.

Q. Recently, RBI has increased the provisioning for unsecured loans and some people are also blaming low wage growth for the rising stress in unsecured loans. So what is your assessment about it?

I think it is fair. Banks want to tap the unsecured loan market as it pumps more money back into the economy. But at some point when there is stress in the system, or when the job market is low, there could be a reversal in that cycle in terms of credit growth. All of that is going to hit you. However, the good news is that unsecured loans form a very small portion of the overall bank's portfolio. So will the banking system be majorly hit by rising delinquencies in unsecured loans? The answer is no.

Q. When there is a credit boom, everybody starts giving the loans and people don't hesitate in taking loans. So instead of mandating provisioning when the crisis unfolds, can’t we track the stress using artificial intelligence?

Artificial intelligence is nothing but simply speaking, non-fashionably speaking, as the past predicts the future. However, it is a power guzzler. We did not have electric power of this magnitude. For example, to run an AI data model. It takes 20 days to compress data and give inputs. So if we are talking about real time and widespread implementation, you need to have enough energy to run a factory. It requires heavy investments in AI infrastructure.

Q. Earlier, Indians used to be known for savings and rarely opt for loans. But now people have no such inhibition for loans. What has changed the people’s attitude?

It is the availability of funds. Your dad could not have bought a house unless he had the complete amount in cash, because banks weren't giving loans. Now you have credit even to go and eat outside. The average consciousness of people in terms of global demands are also different. It's a behavioural shift.

Q. So there have been a lot of incidents relating to cybercrime. So do we have any technological solution for this?

We have an enormous amount of technology available to prevent cyber crimes, but the cybercriminals are also getting smarter. So each time we fix an issue, there is a new issue that comes up. So as much as we are innovating to prevent it, fraudsters are innovating to kind of circumvent it and that is not going to change.