Education Lending: The Risk Below the Surface

Education lending was among the fastest-growing NBFC asset classes for two consecutive years. The thesis was legible: premium foreign universities, an IT employment pathway post-graduation, and overseas income supporting repayment. Three pressures are now compressing that thesis simultaneously. The origination data has already captured much of the impact while the asset quality data has not. The GNPA numbers look clean, but the reason is structural rather than reassuring. The real test has not started yet.

1. The Slowdown Is Already in the Numbers

NBFC education loan AUM grew 48% in FY25 to Rs 64,000 crore, with growth projected to slow to approximately 10-15% in FY26 (as per Crisil). Disbursement growth fell from 50% in FY24 to 8% in FY25. US share in Q1 FY26 disbursements has already dropped to 15% (vs 36% in FY25 and 55% in FY24). Three factors explain what the origination numbers are already showing.

Geopolitical disruption

Since early 2025, a series of US policy moves has targeted international student access. Embassies paused student visa appointments pending expanded social media vetting. OPT (the Optional Practical Training programme) is under active regulatory review. H-1B costs have increased. Canada has separately tightened student visa rules, raising proof-of-funds thresholds and capping permits. Together these moves have undermined the post-graduation employment pathway that is the structural foundation of repayment for US-bound education loans.

Rupee depreciation

The rupee has lost over 10% of its value since the start of 2025. This adds approximately Rs 5-8 lakh per year to the total cost of an overseas degree. Combined with 35-40% tuition cost growth at major destinations between 2021 and 2025, the affordability calculation for the middle-income borrower has shifted materially.

AI and the employment ROI question

IT and engineering graduates form the dominant borrower profile for high-ticket foreign education loans. AI is compressing entry-level hiring in destination markets and in India's IT and BPM sectors. These are the two employment pathways that underpin repayment, whether graduates stay abroad or return. As employment outcomes have become less certain, students and families are recalibrating the ROI of an expensive foreign degree. This is a demand-side pressure that compounds the first two.

2. The Risk the Numbers Don't Show Yet

2.1 The Moratorium Creates a Misleading Picture

NBFC education loan GNPA stood at 0.2% as of June 2025, as per Crisil. The number looks clean, but it is a timing artefact.

Over 85% of the portfolio remains under moratorium. For borrowers already in the repayment phase, the 90+ DPD is approximately 1%, five times the headline figure. The FY23-FY24 cohort, disbursed at peak growth rates, is only now approaching repayment. The DPD clock has not started for the borrowers who matter most.

PSB education loan NPAs have reached 7-8% in prior stress cycles. NBFC portfolios are more selective and have historically produced near-zero NPAs. But those models were calibrated when IT sector hiring was expanding and overseas employment pathways were stable. The ECL and provisioning assumptions built on near-zero DPD data may need revisiting before the FY26-FY27 repayment cohorts arrive.

2.2 The Employment Pathway Is Narrowing on Two Fronts

The FY23-FY24 cohort faces a more specific problem. These borrowers entered loans designed around an employment pathway that has since shifted in two distinct ways.

OPT elimination proposals mean the dollar-income repayment assumption now holds for fewer borrowers than origination models assumed. Loan sizes of Rs 20-50 lakh were structured around overseas earnings. For the cohort already on the books, this is a direct repayment risk.

For the same cohort, the IT employment contraction is another risk factor. This affects both those who remain abroad and those who return to India. India's IT and BPM sector faces the same AI-driven dynamic as destination markets. The Stanford HAI 2026 AI Index documents a nearly 20% decline in employment for software developers aged 22-25.

3. The Adaptive Response Towards Domestic Lending

Geography diversification is already underway. Non-US, non-Canada geographies rose from 25% to approximately 50% of new disbursements in FY25. NBFCs have built active university partnerships across the UK, Ireland, Australia, Germany, and New Zealand. Domestic lending is the longer-term structural move, and genuine demand tailwinds support it. 

India's GCC ecosystem employed approximately 2 million professionals by end-2025. It is on track for 2.5-2.8 million by 2030, generating demand for specialised technical education. The PM Vidyalakshmi scheme has expanded collateral-free domestic loan access. It offers a 3% interest subsidy on loans up to Rs 10 lakh for eligible borrowers. 

However, given lower domestic loan ticket sizes, their share of total AUM will remain limited near-term. The domestic pivot stabilises the book but does not substitute for foreign-study volume. Also, executing it at scale requires a different product architecture than what most education lending platforms were built for.

4. The Domestic Pivot Requires a Different Operating Model

Executing the pivot is not a credit policy decision. It requires genuinely different product architecture.

• Collateral structure. Loans below Rs 7.5 lakh are unsecured, covered by the CGFSEL government scheme. Above that threshold, tangible collateral is required at banks. NBFCs can lend unsecured above this level for premium-institution admissions, but the underwriting logic changes entirely. In foreign lending, the admission offer and expected earnings largely substitute for collateral.

  
  

• Disbursement mechanics. Domestic loans disburse in tranches, semester by semester, paid directly to the institution. Foreign loans typically disburse in one or two tranches. The LOS must handle tranche-level schedules with institution-linked release conditions.

  
  

• Underwriting logic. Institution rank drives pricing explicitly. Several lenders offer rate concessions for top-20 NIRF-ranked institutions. This is distinct from the admission-plus-employment-pathway model used in foreign lending.

  
  

• Servicing timeline. Domestic PG and MBA programmes run 1-2 years versus 2-4 years for foreign degrees. Moratoriums are shorter. Repayment begins earlier. LMS servicing configuration differs accordingly.

  
  

• Bureau depth. Tier 2/3 city borrowers carry thinner credit files. Alternative data and co-applicant assessment carry more weight than in the premium-institution foreign study model.

Running a meaningful domestic book alongside an existing foreign book is not a product variant exercise. It requires two product architectures within the same platform.

5. How OneFin Supports This Transition

OneFin's unified LOS/LMS platform is used by education lending NBFCs managing this transition. The relevance is specific to the operational gaps the domestic pivot creates.

(a) Tranche-level disbursement management. OneFin supports configurable disbursement schedules at tranche level for domestic loans. Single-disbursement processing for foreign loans runs within the same platform. Institution-linked release conditions can be configured without custom development.

(b) Secured lending workflow support. Domestic loans above the Rs 7.5 lakh threshold require property valuation, legal verification, and mortgage documentation. These workflows are available in OneFin's platform and are not part of a standard foreign-study origination process.

(c) Fixed-rate and floating-rate product support.OneFin supports both fixed-rate and floating-rate loan structures within the same platform. This includes configurations with and without moratorium periods, across both foreign study and domestic products.

(d) Dual-book LMS servicing. A single LMS can be configured to manage long-moratorium foreign loans and shorter-cycle domestic products concurrently. This reduces the platform fragmentation that typically follows a portfolio pivot.

(e) Bureau and thin-file underwriting. API-level bureau integration supports alternative data inputs for Tier 2/3 domestic borrowers. Co-applicant assessment and alternative scoring carry more weight than bureau depth for these profiles.

Conclusion

The structural demand for education financing in India is intact. The sector is still growing. The shift toward new geographies and domestic products is already underway across most large NBFCs. What is less certain is the asset quality picture waiting on the other side of the moratorium. 

The FY26-FY27 repayment window will be the first real test. It arrives for a cohort disbursed at peak growth, entering employment under conditions no origination model from that period anticipated. Geography diversification and the domestic pivot are the right responses. Both require operational infrastructure built for genuinely different product architectures, not just different credit policies. The institutions recalibrating now will be better placed when the test arrives.

To know more about OneFin, schedule a Demo.