Banking Sector Shows Resilience Despite Unsecured Loan Stress: Report
GNPAs surged from 3.8 per cent in FY14 to a peak of 11.2 per cent in FY18, while net NPAs more than doubled from 2.1 per cent to 5.6 per cent over the same period
You're reading Entrepreneur India, an international franchise of Entrepreneur Media.
India's banking sector has sustained a broadly benign asset quality through FY25, despite emerging stress pockets, especially in unsecured personal lending. The sector's gross non-performing asset (GNPA) ratio stood at 2.3 per cent at the end of Q4 FY25, marking a significant turnaround from the post-AQR years that saw surging defaults and a systemic clean-up in the aftermath.
This improvement in asset quality is the culmination of a decade-long recalibration that began with the Reserve Bank of India's asset quality review (AQR) in 2015-16. Banks were then compelled to recognise bad loans and limit excessive restructuring practices, which revealed deep cracks in wholesale lending portfolios. As per a recent report by CareEdge Ratings, GNPAs surged from 3.8 per cent in FY14 to a peak of 11.2 per cent in FY18, while net NPAs more than doubled from 2.1 per cent to 5.6 per cent over the same period. In response, banks shifted focus sharply toward retail lending, expanding it from 19 per cent of advances in 2015 to 34 per cent in 2025—seeking safety in smaller, diversified exposures.
This pivot helped banks limit further slippages, aided by steady provisioning, aggressive write-offs, and recoveries. The trend continued through the pandemic years, buffered by regulatory forbearance and temporary moratoriums, and was further entrenched in FY25. Private sector banks (PVBs), however, have shown higher slippage ratios than their public sector counterparts, largely due to a heavier tilt toward unsecured retail credit. In response, scheduled commercial banks (SCBs) not only increased write-offs but also continued to sell stressed assets to asset reconstruction companies (ARCs), steadily cleaning their balance sheets.
Still, emerging risks cannot be ignored. Sanjay Agarwal, senior director at CareEdge Ratings, pointed to rising vulnerabilities in the personal loans segment. "Net additions to NPAs have remained broadly low, enabling the sector to witness a steady reduction in headline asset quality numbers. However, with the personal loans segment facing stress, the overall fresh slippages are expected to rise, and recoveries/upgrades are likely to taper gradually," Agarwal noted. He projected a marginal deterioration in GNPA ratios, estimating a shift from 2.3 per cent in FY25 to between 2.3 per cent and 2.4 per cent by FY26. "Key downside risks include deteriorating asset quality from elevated interest rates, regulatory changes, and global headwinds such as tariff increases," he warned.
On the policy front, the Reserve Bank of India's unexpected 50bps rate cut in FY25, along with a phased 100bps CRR reduction set to roll out from September 2025, has added fresh momentum to the banking landscape. Naveen Kulkarni, chief investment officer at Axis Securities PMS, described the move as a surprise shift from the expected gradual easing. "The regulator is in favour of front-loading rate cuts to support growth. However, the RBI has changed its stance from Accommodative to Neutral. This provides limited scope for further rate cuts," he said.
Kulkarni added that the regulator's downward revision of the FY26 inflation forecast to 3.7 per cent and its maintenance of the GDP growth estimate at 6.5 per cent signal cautious optimism. He emphasized the importance of liquidity support for banks as they aim to revive credit growth in the second half of FY26. "Asset quality concern appears to be steadily waning with unsecured segment stress showing gradual signs of stability, while the secured segment asset quality continues to hold up well," Kulkarni observed.
While the systemic GNPA numbers remain low, the coming months will test whether the stability holds as unsecured lending continues to expand and macroeconomic pressures play out.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Bloomberg
an hour ago
- Bloomberg
Sam Altman Chats Virtual Companions at Mizuho Tech Conference
Hi, it's Liana Baker in New York. Today, I'm bringing you highlights from a banker interview with OpenAI CEO Sam Altman that I sat in on. Elsewhere, Adnoc makes one of its most audacious overseas moves to date. Today's top stories
CNET
2 hours ago
- CNET
Wednesday's Fed Decision Could Actually Help Boost Your Savings. Here's How
High interest rates can benefit savers. Getty Images/Amy Kim/CNET A few times a year, you probably see headlines about the Federal Reserve's interest rate decisions. It may seem like boring economic news, but it matters more than you think. The Federal Open Market Committee, the central bank's decision-making body, meets throughout the year to determine monetary policy and adjust the federal funds rate, which affects the cost of lending between banks. When interest rates are high, like they are now, the banks pay more to borrow money, and so do you, for loans, mortgages, credit card APRs, etc. Also, when interest rates are high, some banks offer better returns for the money and deposits you keep in their institutions. The Fed doesn't directly set what banks charge for loans or pay on savings accounts, but it tends to have a ripple effect for customers. Once the Fed cuts the benchmark rate, we can expect cheaper loan rates and lower returns on your investments. In short, the Fed doesn't just impact borrowing costs but savings too. Higher interest rates are good if you're trying to score a higher yield on your earnings and investments. You'll just need the right kind of account. The Fed's next decision will be announced on Wednesday. Here's how it will impact your hard-earned cash. How the federal funds rate impacts your money Raising the federal funds rate is one tool the Fed can use to help tame inflation. When consumer prices are high, the central bank hikes interest rates to discourage borrowing and to slow the economy. That's the pattern we saw between early 2022 and 2023. When inflation is more or less under control but unemployment is high, the Fed cuts interest rates to make it easier — and more enticing — for households to spend money. The Fed carried out three rate cuts in late 2024. We've all been waiting for more rate cuts, but the Fed has held rates steady so far this year to see how Trump's policies, particularly tariffs, affect the economy. Right now, markets say the Fed will keep the federal funds rate range the same at its June 17-18 meeting. The Fed isn't likely to cut interest rates again until September. When those rate reductions come later this year, except to earn a lower annual percentage yield on certain savings accounts. Holding interest rates high gives your savings a boost Before you make any financial moves this week, take a moment to assess your savings strategy. First, look at where you're parking your money. Traditional savings accounts with brick-and-mortar banks offer the lowest returns on your savings, 0.02% at best. These accounts are often paired with your checking account, an obvious convenience if you prefer to keep all of your money with one bank. However, if you want to take advantage of today's high-interest rate environment, I recommend a high-yield savings account. Online-only banks and credit unions usually offer high-yield savings accounts, some with APYs between 3% and 4%. Those rates are still variable, so they could go down in a few months. But you'll still make more money back with an attractive rate than you would at a traditional bank. If you're not convinced, here's a comparison chart: Comparing savings accounts Traditional savings account High-yield savings account Mostly brick-and-mortar, so you can do transactions in person. Mostly online, so you won't visit a physical branch too often. APYs around 0.02%. If you have $1,000 in savings, you'll earn a couple of dollars after a year. Best APYs around 4%. If you have $1,000 in savings, you could earn $30 to $40 after a year. Often comes with maintenance fees that can dig into your savings. Usually has lower or no fees because there are fewer overhead costs.
Forbes
2 hours ago
- Forbes
Agentic AI In Enterprise QA: Powering Intelligent, Autonomous Testing At Scale
Pradeep Govindasamy is the Co-Founder, President and CEO of QualiZeal. We're at the beginning of a new era in quality engineering, one shaped by agentic AI. While generative AI has captured global attention, the real transformation in software testing is only just beginning. I believe we're now entering a phase where AI isn't just assisting people in testing tasks. It's becoming autonomous, goal-driven and capable of acting with intelligence across the lifecycle. At QualiZeal, we're witnessing this shift firsthand. As someone who has spent years in the testing space, I can confidently say that AI is not a far-off future. It's here, being built into our processes today, and it's already beginning to disrupt how we think about quality at scale. Software development and testing are the two most critical pillars in any IT application lifecycle. To get a product into the hands of customers, you first build it, then test it and only then can you ship it. We've seen how tools like GitHub Copilot have revolutionized development. Now, that same level of AI adoption is happening in software testing. This is no small market—it's a $100 billion global industry. And just as smartphones once disrupted legacy devices like BlackBerry, AI is poised to transform testing in a similar way. Every phase of the software testing lifecycle—test case preparation, test design, test data management, performance testing, site reliability engineering—is now being infused with AI to increase efficiency, productivity, and ultimately software quality. Before we talk about agentic AI, we need to understand the evolution. The first step in embracing AI is automating repetitive, rule-based tasks. Once you have robust automation in place, AI capabilities can be layered on top to improve every phase of testing. But agentic AI goes one step further. With standard AI, we build prompts, define logic and teach the models how to behave. With agentic AI, we create systems that learn, adapt and act autonomously. These agents follow instructions and understand intent. They can analyze changes in the system, adjust automation scripts accordingly and execute tests without human intervention. For example, imagine a scenario where a company updates its checkout process, maybe tariffs or payment options change. In the past, a QA team would have to manually identify changes, rewrite test scripts and rerun tests. With agentic AI, the system learns what's changed, modifies the scripts, self-heals when errors occur and continues testing. It even generates a report outlining what it changed and why. This self-healing, self-optimizing capability sets agentic AI apart from traditional automation. And it's a game-changer. We're seeing both technical benefits and measurable business outcomes. With agentic AI, the cost of quality is decreasing. From my observation, the industry average today is about 18%, but with AI-infused testing, we anticipate a 5% drop, driven by reduced manual effort and increased efficiency. In maintenance alone, we've seen a reduction from 20% of team capacity to less than 5%. Even more importantly, release cycles are accelerating. Time to market (TTM) has gone from quarterly to weekly, and now, with agentic AI and DevOps practices, to daily releases. The entire production throughput is becoming faster and more reliable. And decision-making is more seamless because agentic systems provide full transparency through real-time reporting, eliminating the need to compile data across disparate systems. Organizations looking to lead in this space must prepare now. I always say this moment is not just about catching up—it's about disrupting yourself before you get disrupted. Companies that wait too long will miss the opportunity to lead. Those who invest now will be in a position to capture market share and build the next generation of testing capabilities. This preparation requires both a top-down and bottom-up approach. Leadership must allocate budgets, not just wait for client-driven funding, and teams must be empowered to get trained, certified, and exposed to different AI models. AI isn't just a CIO or CTO conversation anymore. It's happening at the board level, and for good reason: this is the foundation for long-term competitiveness. I recommend organizations push their teams to reach at least level three in AI readiness: basic execution. Core functions like engineering and QA need to go further, while ancillary teams like finance and marketing should also gain exposure. Of course, with great power comes responsibility. We need to ensure agentic systems operate ethically, transparently and securely. Especially in regulated industries like healthcare, insurance or banking, any AI-driven decision, no matter how small, can have massive consequences. That's why testing the AI itself is just as important as using AI for testing. There's a growing demand for AI-specific test engineers who can validate agentic systems through high-end exploratory techniques. Traditional testing models like equivalence partitioning or boundary analysis must now be complemented with new approaches tailored to AI behavior. In the near future, eight to 10 new job roles will emerge specifically to test and validate agentic AI systems. These won't be optional. They'll be mission-critical. We estimate that full-scale AI maturity across the testing lifecycle will arrive around 2027. Between now and then, we're in the planning and education phase, training models, customizing LLMs and building the necessary infrastructure. Implementation will accelerate in 2026, and by mid-2027, I expect the majority of enterprise QA environments to be agentic by design. This is a once-in-a-generation opportunity for testers, developers and technology leaders. Gen Z professionals, especially those raised in a digital-native world, will have an edge. They can adopt these tools faster, and many will find themselves building careers in entirely new domains. We're not just building testing systems anymore. We're building trusting systems. Platforms that learn, adapt and support business continuity without human babysitting. That's the future of QA. That's where agentic AI takes us. And the companies that embrace it today? They'll be the ones defining quality tomorrow. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
