The Financial Stability Report of July 2020, by the Reserve Bank of India (RBI), indicates that the gross NPA ratio of the scheduled commercial banks can rise from 8.5% in March to 12.5% in March 2021.
Although Non Performing Assets (NPAs) have been written off by the financial institutions time and again, they are still the top concern for them. An upward trajectory of NPA is adding to the woes for India’s economy in recent years, threatening the legitimate economic output, productivity, and employment. Expanding the credit portfolio while managing NPAs and retaining good revenues will be the key challenge for Indian Banks and NBFCs in the future as the lending market, and the banking sector is under tremendous pressure to address the rising NPAs in the country. Here, RBI’s stringent policies have come as much-needed support for the lending community, amidst the ordeal. Yet, a significant shift towards tech-driven innovative tools of the trade will broadly drive the change.
The Policy Intervention
RBI issued Master Directions for compliance to private and public banks and some select financial institutions in 2017 for monitoring their respective portfolios using the Early Warning Signals (EWS) Framework. RBI and the Department of Financial Services (DFS) set a thorough framework, listing 42 and 83 signals, respectively, to be produced on the data obtained from these sources. Thus, automating the monitoring of a borrowing entity’s health in real-time is now mandatory for all lenders.
Thrust from government authorities to share regulatory, tax, and other pertinent data vaults to the public have implemented EWS possible in the current banking landscape. Moreover, a swiftly changing tech ecosystem has enabled ML-driven algorithms to comb through huge structured and unstructured datasets. A 24*7 on-alert allows Machine Learning algorithms to identify signs of deteriorating economic conditions of borrowers, classifying them as Red Flagged Accounts (RFAs).
The Role of AI and ML technologies
Most Indian enterprises have sparse data points. And more than 60 million MSMEs have a negligible digital footprint. Monitoring them leveraging traditional data points will not be sufficient to assess their risk factor. Hence, a 360-degree view of the borrower’s information from previously untapped or alternate data repositories becomes imperative. This additional data ranges from credit card usage patterns, bank cash flow analysis to non-banking inputs such as mobile or Wi-Fi bills, utility bills, etc., of such businesses.
Such data sources may not seem relevant in the first place but provide deep insights about the responsible attitude of a person towards economic commitments. It works the same way that good academic background and certifications indicate an employee’s potential. An automated EWS assesses all these data points, eliminating the threats of bias, high costs, and errors of manually estimating the creditworthiness of long-tail MSMEs.
- The system picks prescribed signals of distress 12-18 months in advance even before the occurrence of the threat.
- They are used for real-time fraud identification, improved compliance, predicting frauds, and taking preventive steps, leveraging versatile datasets.
- AI and ML can also work towards NPA resolution. The recent RBI guidelines on NPA resolution focus on the need for a ‘resolution plan’ for accounts in default. Lending entities can use ML tools to build relevant and specific resolution plans that would have a higher chance of success by learning from patterns in resolution based on historical strategies, customer segments, product types, among other factors. Collection efficiencies can be streamlined by offering actionable insights to collection channels on successful strategies.
- AI/ML algorithms can detect and prevent suspicious transactions, empowering a leap from defective to preventive credit risk monitoring.
- Natural Language Processing (NLP) models process unstructured data such as social media, news, and unconventional data for sentiment analysis.
- With time, as algorithms improve, ML-driven EWS suppresses false-positives and sets thresholds at optimal levels, thereby issuing more accurate results.
Huge datasets, automation, and process augmentation can bring more transparency in auditing. This will also mitigate apprehensions around responsibility from bankers later in the life-cycle of the loan account. A robust policy riding on the back of analytical technologies in risk management solutions will take the lead role in reducing the issue of NPAs and boosting growth in the economy.