MSME NPAs Mirroring Trends in Overall Banking Sector: CGTMSE CEO Manish Sinha

Collateral-free loans have been a lifeline for India’s micro, small, and medium enterprises. But behind recent credit growth to MSMEs lies a simmering churn within the Credit Guarantee Fund Trust for Micro and Small Enterprises (CGTMSE) – the institution backing such loans to micro and small enterprises (MSEs). Speaking with FE Aspire, CEO Manish Sinha describes the change that has enabled it to cross the 1 crore guarantee milestone—more than half of which was issued in the last four years.

From streamlining oversight and reducing charges to exploring AI-based credit frameworks, Sinha describes how CGTMSE is seeking to remain relevant in an evolving credit environment. He also responds to concerns regarding accountability at the borrower level, NPA trends, co-lending, and more. Edited Excerpts below: 

In the last 3–4 years, CGTMSE has issued more guarantees than it did in the previous two decades combined. What triggered this turnaround?

From 2000 to 2021, we issued around 52 lakh guarantees. But in the last three to four years alone, we have issued more than that, taking us past the 1 crore mark. In value terms, the total guarantee amount jumped from Rs 2.66 lakh crore (till FY21) to Rs 9.35 lakh crore cumulatively by FY25. In FY25 alone, we issued guarantees worth Rs 3.05 lakh crore—our highest in any single year. This year, we are aiming for Rs 4-4.5 lakh crore.

This scale-up was the result of multiple strategic shifts. We began by identifying key hurdles—whether in policy, process, or systems. With government support, we secured additional funds, made systemic reforms, and implemented important policy changes. 

One of our most significant efforts was raising awareness. We conducted over 100 seminars and workshops across the country, not just for enterprises but also for bankers because we found that many were unaware or unsure about our schemes. We also engaged with industry associations and participated in third-party events to expand our outreach. 

We then looked inward. We removed redundant requirements, relaxed norms, and offered special provisions—for example, enhancing coverage for women-led enterprises to 90 per cent and easing norms in credit-deficient districts. The guarantee ceiling was raised from Rs 2 crore to Rs 5 crore, and we rationalised the guarantee fee structure, bringing it down to as low as 0.37 per cent. 

We also collaborated with state governments and improved our technology infrastructure to streamline operations. These changes—internal and external—came together to help us scale in a sustainable manner. We're also working on a few new digital products that are more relevant to bankers in the current scenario. 

What are these new products? 

It is too early to talk about them. They are still on the drawing board. Today, banks drawing data from the available public data sources are creating products around digital lending. This is one area we feel going ahead will be a major source of credit to MSEs. CGTMSE certainly sees a space that it can fill up. Cash flow-based lending is another thing that we are looking at. 

We are also exploring whether it's the only guarantee that we need to give on loans or there could be something else in addition to the guarantee. So, some new products will definitely be coming out this financial year. 

What are the current NPA (non-performing asset) trends in CGTMSE-backed loans? 

We closely track what’s happening both within and outside the CGTMSE portfolio. You would be happy to note that there’s no major discrepancy between the two. The trends in NPA levels in CGTMSE-covered loans broadly mirror the overall lending landscape. 

This shows that banks are offering credit guarantees to a wide range of MSEs, not just high-risk borrowers. If guarantees were being extended only to poor-quality borrowers, our NPA levels would be disproportionately high, which is not the case. So, banks are taking a balanced and inclusive approach under the CGTMSE scheme. 

There’s also a common misconception that when a 90–95 per cent guarantee is available, banks might skip due diligence and lend indiscriminately. I’d like to clarify that this is incorrect. What CGTMSE provides is coverage of the loss-given-default (LGD), not protection from the probability of default (PD). 

If the probability of default is not changing, no bank would lend just because the CGTMSE guarantee is available. After the credit guarantee is taken care of, the net NPA can come down. But what happens to the gross NPA? Gross NPA will show a very high figure, and anyone can make out that the banker is not doing a good job. So, the gross NPA has to be controlled. 

Overall, can you share the NPA figure? 

We don’t have a precise figure for net NPA at CGTMSE; however, like I said, our observations mirror trends seen in the overall banking sector and there’s no significant deviation. It also depends on the provisioning done by individual lenders. Segment-wise, we have noticed that NPAs tend to be slightly higher among NBFCs, while public and private sector banks show relatively lower levels.  

We regularly monitor these trends to ensure that lenders aren’t offloading only their stressed assets into our portfolio. We compare the NPA levels reported by banks and NBFCs publicly with those seen in the loan portfolios they place with us. So far, we haven’t seen figures going haywire, but we continue to keep a close watch. If needed, we are prepared to step in and take corrective measures to safeguard the scheme’s integrity. 

How has the scheme performed in co-lending loans?

Co-lending is a growing area for us. We track emerging market trends and step in where needed. Co-lending was one such opportunity. The response has been encouraging so far. In FY24, about 932 MSEs were disbursed Rs 244 crore loan through co-lending. This jumped to 6,202 MSEs receiving Rs 1,499 crore in FY25. For a new product, that’s a great start. However, there are still several operational and structural challenges that lending institutions need to address to scale the co-lending model further for MSEs. 

A Niti Aayog report in May this year called for bringing CGTMSE under a stronger regulatory authority, citing a lack of oversight as a challenge to balancing fund availability with financial discipline. What’s your take? 

I wouldn’t like to comment on that, but I will give you the world perspective. The institutions providing credit guarantee in other countries in a regulatory environment are the ones that are also into lending. We are not into lending. We facilitate lending via banks to help MSEs raise formal credit. That said, CGTMSE is also subject to multiple layers of audit. We undergo audits by the CAG, statutory audits, and internal audits. 

Our partner institutions—banks and NBFCs— to whom we provide guarantees are all regulated by the RBI. SIDBI, from whom we get funds, is also subject to regulations. However, if the government decides on any regulatory change, of course, we have to follow that.

Have you seen cases where borrowers feel less accountable after getting a collateral-free loan? How do you ensure that collateral-free doesn't become accountability-free?

That’s an important concern. While CGTMSE offers a guarantee, the responsibility of monitoring, recovery, and follow-up lies with the banks. While there is no collateral involved, primary security has to be there with the normal recovery processes. If we find a bank or NBFC’s portfolio underperforming or not aligned with market norms, we impose a risk premium on them. This system works well as lenders strive to avoid the additional cost. While we expect some NPAs but banks not following the normal process of monitoring and recovery just because the government guarantee exists is not acceptable.

The government last year revised the CGTMSE target with an additional Rs 5 lakh crore guarantee in two years. What are your plans for it? 

Our focus is on multiple fronts. First, as I said earlier, awareness remains critical among MSEs across sectors, and clusters still lack awareness. We're addressing this through structured outreach and education. Second, we're working closely with bankers. Continuous engagement and training are essential, especially in case of staff transfers, the new people have to be made aware of credit guarantees. Some banks have helped by organising sessions for their credit officers, and we aim to scale this further.

Third, we’re expanding partnerships. All Regional Rural Banks (RRBs) are now eligible for CGTMSE, and we are also onboarding NBFCs and cooperative banks to broaden our reach, especially in underserved and remote areas. There are around 300 lending partners, including all public and private sector banks, most RRBs, cooperative banks, and a significant number of NBFCs. We are also actively seeking to onboard more partners to deepen credit access, especially in underserved regions. 

What about product and technology improvement? 

On the product side, a major focus is on green and sustainable lending in areas like solar energy and energy-efficient machinery. New guarantee mechanisms are being considered to make credit more accessible for MSEs adopting such technologies. 

In parallel, CGTMSE is also improving technology to streamline processes further. While guarantees are currently issued without manual intervention for many banks through APIs, the aim is to extend this seamless integration to other partners as well. This will enhance speed, efficiency, and ease of operations for lenders. Broadly, CGTMSE’s strategy revolves around four pillars – product, process, technology, and awareness.  

With AI-driven lending on the rise, what role will trust-based mechanisms like CGTMSE play? Will it need to adopt a more data-integrated model? 

Absolutely. That’s one of my dream projects. With the pace at which AI is advancing, we can’t ignore it. We sit on a huge volume of data, and AI will increasingly guide our decision-making, just as it will for banks. We're already working on integrating AI to assess geographies and industries more effectively. It’s essential for CGTMSE to evolve into a data-driven, tech-integrated model.