A credit risk management in NBFCs tells you how the portfolio looks today. Vintage analysis tells you how the credit decisions made 18 months ago are performing now and whether the decisions being made today will look better or worse in 18 months.
Portfolio vintage analysis for NBFCs is the analytical framework that separates the quality of current decisions from the performance of historical decisions. It is the most rigorous way to evaluate whether an NBFC’s underwriting is improving, deteriorating, or stable over time.
What Portfolio Vintage Analysis Is and Why It Matters for NBFCs
Portfolio vintage analysis groups loans by the month or quarter they were disbursed, the vintage, and the DPD in credit reports rate over time from origination.
A vintage curve shows: for loans disbursed in Month X, what percentage had reached DPD 30+ by Month 3, Month 6, Month 12, and Month 24 of their life?
The power of vintage analysis is that it separates the causes of current portfolio performance. An NPA classification in India today could reflect good old loans going bad (origination quality was fine, external stress is hitting seasoned loans) or bad new loans turning delinquent early (origination quality has deteriorated). Vintage analysis distinguishes between these two entirely different situations.
How to Build a Vintage Curve for an NBFC Loan Portfolio
The construction of a vintage curve requires four data elements per loan:
- Disbursement date (to assign the loan to a vintage cohort).
- bank statement analysis for NBFCs.
- Current DPD status at each observation month.
- First DPD 30+ date (to identify when each loan first became delinquent).
The vintage curve for a single cohort, say, all loans disbursed in Q2 2024, tracks the cumulative delinquency rate of that cohort at each month of loan life:
- Month 3: What percentage of Q2 2024 loans had reached DPD 30+ by their third month?
- Month 6: cumulative delinquency at six months.
- Month 12: cumulative delinquency at 12 months.
- Month 24: cumulative delinquency at 24 months.
Plotting multiple cohorts on the same chart, Q1 2023, Q2 2023, Q3 2023, Q4 2023, Q1 2024, produces the vintage curve stack that is the primary analytical output.
Reading Vintage Curves: What Divergence Means
Vintage curve patterns reveal specific origination quality and portfolio performance stories.
Parallel Cohort Curves (Stable Quality)
When multiple vintage cohorts track closely together Q1 2023, Q2 2023, and Q3 2023 all showing similar cumulative delinquency rates at each month of life, the origination quality is stable. The NBFC is consistently selecting borrowers of similar credit quality across different time periods.
Rising Newer Cohort Curves (Deteriorating Quality)
When newer cohorts are tracking worse than older cohorts at the same age, the Q4 2024 cohort shows 3% cumulative DPD 30+ at Month 6, while Q4 2023 showed 1.5% at the same age; origination quality has deteriorated. The newer loans are performing worse than older loans did at the same point in their life.
This pattern is the most important signal in vintage analysis for NBFCs. It means the current NPA rate understates the future NPA rate; the loans being made today are of lower quality than the loans that make up the portfolio now.
Improved Newer Cohort Curves (Improving Quality)
When newer cohorts track below older cohorts at the same age, origination quality is improving. The newer loans are performing better than older loans did at the same life stage. This typically follows a credit policy tightening, an improvement in automated underwriting accuracy, or a change in product mix toward lower-risk borrower segments.
Using Vintage Analysis to Identify Underwriting Changes
Portfolio vintage analysis is the most reliable tool for identifying when and how underwriting changes have affected portfolio quality.
Changes in origination quality typically appear in vintage curves 3 to 6 months after the change occurs because early delinquency risk materialises earliest at Months 3 to 6 of loan life. A credit policy change in April 2025 will first be visible in the divergence between the April 2025 cohort and the March 2025 cohort at the Month 3 observation point (July 2025).
This means NBFC credit risk teams that monitor vintage curves monthly, not just quarterly, can identify origination quality changes 9 to 15 months before they would become visible in the NPA ratio. That early identification window is the risk management value of vintage analysis.
Vintage Analysis in RBI Regulatory Reporting and Stress Testing
Portfolio vintage analysis has become an expectation in the RBI examination of Middle Layer and Upper Layer NBFCs. RBI examiners and supervisors increasingly ask credit risk teams to demonstrate that they understand which cohorts are performing well and which are under stress, not just what the aggregate NPA ratio is.
For stress testing requirements (applicable to NBFC-ML and NBFC-UL), vintage analysis provides the empirical basis for default rate assumptions under adverse scenarios. Historical vintage curves show how quickly delinquency accumulated in previous stress periods, providing calibrated inputs for forward-looking stress scenarios rather than arbitrary assumptions.
Key Takeaways
- Portfolio vintage analysis for NBFCs groups loans by disbursement month and tracks cumulative delinquency rates over the loan life, revealing whether origination quality is stable, deteriorating, or improving.
- Rising newer cohort curves, where loans disbursed recently are tracking worse than older loans at the same age, are the most important warning signal in vintage analysis. It means current NPA understates future NPA.
- Vintage divergence typically becomes visible 3 to 6 months after an underwriting quality change, enabling identification 9 to 15 months before the change appears in the aggregate NPA ratio.
- RBI examiners for NBFC-ML and NBFC-UL increasingly expect vintage analysis as part of credit risk monitoring, not just aggregate portfolio metrics.
- Historical vintage curves provide empirical default rate inputs for stress testing, replacing arbitrary scenario assumptions with observed portfolio behaviour under previous stress conditions.
Frequently Asked Questions
What is vintage analysis in the context of NBFC credit risk management? Vintage analysis groups loans by their disbursement month or quarter (the “vintage”) and tracks the cumulative delinquency rate of each cohort at each month of loan life. Plotting multiple cohorts together produces vintage curves that reveal whether origination quality is stable, improving, or deteriorating across different time periods.
How does vintage analysis differ from standard NPA monitoring? Standard NPA monitoring shows the current percentage of the total portfolio that is non-performing, a single aggregate number that reflects all historical origination decisions. Vintage analysis shows how each cohort of loans originated in a specific period is performing at each stage of its life. This separates the performance of recent origination decisions from the performance of older loans, revealing whether current underwriting quality is better or worse than historical underwriting quality.
When does a credit policy change become visible in vintage analysis? A credit policy change typically becomes visible in vintage curves 3 to 6 months after the change is implemented, corresponding to the earliest delinquency risk period (Months 3 to 6) for loans originated under the new policy. A credit policy tightening in April 2025 will show improved performance in the April 2025+ cohorts compared to Q1 2025 cohorts, visible at Month 3 to Month 6 observation points from July 2025 onward.
Do RBI examiners expect NBFC-ML and NBFC-UL entities to conduct vintage analysis? Yes. RBI supervisory expectations for NBFC Middle Layer and Upper Layer entities include portfolio-level risk monitoring that goes beyond aggregate NPA ratios. Vintage analysis by origination cohort is increasingly an expected component of the Credit Risk Committee reporting and the annual ICAAP (Internal Capital Adequacy Assessment Process) for Upper Layer NBFCs.
What data does an NBFC need to build vintage curves? The minimum data requirements are: loan disbursement date, loan amount at disbursement, monthly DPD status for each loan from origination to the current date, and first DPD 30+ date for each defaulted loan. Most loan management systems contain this data; the analytical challenge is structuring it into cohort-based cumulative delinquency calculations rather than point-in-time snapshots.
Conclusion
Portfolio vintage analysis is the credit risk management tool that asks the most important question in NBFC lending: which loans made last year are the most important?
The aggregate NPA ratio answers a different question: how are all loans performing today? Vintage analysis answers the prospective question, and that prospective view is what separates reactive credit risk management from proactive credit risk management.
Build the vintage curves. Read them monthly. Act on the divergences before the next RBI examination asks you why you missed them.
July 20, 2026
7 min read
