Advanced Risk Analytics Mitigates The Bad Loan Problem

Capital is a scarce resource in India. Banks’ bad loan problem is only making it worse. More than Rs 9.5 lakh crore of funds is tied up in bad loans. Such capital is lost to unproductivity, which further curtails capital available that could have been better used in productive assets and investments.

A bigger problem is banks have to be recapitalised which means that more capital is needed to fix banks’ broken balance sheets, hence more resources have to be diverted to banks. So bad loans not only drives a wedge in a bank’s balance sheet, it also constricts economic expansion.
The bad loan malaise has been widespread and not restricted to just the public sector banks. Private sector banks are at the receiving end too with supposedly better models to assess risks. Big ticket loans have gone bad despite financial institutions of high repute being a part of a consortium that took financial exposure.

There are two issues at stake here. One, has this problem ballooned due to an internal lapse in the requisite due diligence on the borrower. Two, business losses arising out of market conditions are difficult to assess beforehand leading to inferior lending decisions.

In fact, NPAs are the result of two key components:

1.    Lapses in evaluation of borrower’s risk profile (Pre-Facto)
2.    Efficiency and effectiveness of the recovery mechanism adopted by the financial institution (Post-Facto)
Both these drawbacks can be hugely mitigated through investments in advanced analytics and data tools.

Analytics in credit worthiness
New lending firms that using analytics and artificial intelligence to predict credit behavior are at an advantage. The critical issue is to gather as many data points and build sophisticated lending models that can analyse vast array of data sets to determine how and when an individual/business is likely to default.

Data is complex task. Often the information is incomplete or misleading. Hence, data analytic models have to integrate various platforms of data to cross-verify and authenticate data from different sources. Accuracy in lending can be further built upon using widely back-tested proprietory tools and models to further strengthen the lending mechanism.

Digital footprints leave valuable information  
New-age financial firms have started evaluating an individual’s vast digital footprint. An individual’s purchase history, app history, search history are some further tools that are being increasingly used by the new age lending firms to determine creditworthiness.

More people spend more time digitally. Using various analytic tools, and inbuilt risk-assessment models a fairly accurate assessment can be made on how an individual is likely to behave on his borrowing. AI-based models can scour tons of data and make the job of risk-assessment a lot easier.

Financial firms that use an automated system to monitor behaviour of the sectors and industries are also at an advantage. The rapidly changing eco-system means that there are several imbalances in the system that need to be constantly monitored once a loan is disbursed. Even if the loans are good, individuals and businesses have to be sliced and diced on a regular basis to estimate pain point that could emerge in the repayment cycle.

Keep tabs at low cost
Data analytics is the big-game changer on this front. Several non-banking financial firms have put up several risk assessment metrics that can help to identify potential bad loans at a very early stage.

The system gathers and integrates data from both internal systems as well as a variety of external sources - like the web, external credit rating agencies, regulatory & legal bodies, industry associations, customer’s spending pattern, etc. It uses this data to derive useful information to monitor and analyse customer behaviour.

By applying real time big data analytics and predictive analytics to extract actionable intelligent insights and quantifiable predictions, financial firms gains insights that encompass all types of customer risk profiles.

For example, defaults in other loans taken is a significant indicator of potentially fraudulent action, and yet, many solutions either ignore or are unable to capture and react to these indicators.
Firms that have an expertise in deploying this mix of technology, analytics and human-centric knowledge provides necessary information to take pro-active measure which can hugely reduce the risks of defaults.

AI-led lending to be a game changer
The use of data analytics and machine learning to process many sources of data to determine credit worthiness is growing significantly. The more data you gather, the more better one would be able to predict repayment behaviour, including the chances of a likely default.

Data by itself is meaningless. Firms need to make data talk and reveal insights that more accurately predict the chances of default. New AI led-lending companies have that predictive platform that can change the lending landscape, and the more banks and NBFCs use it, the better it is for the lending business.