Aggregate NPA ratios tell you how bad things are right now. Vintage analysis tells you when things started going wrong, which loans are responsible, and what underwriting or market conditions created the problem. For any NBFC managing a loan portfolio, these two types of information are complementary and vintage analysis is the one that enables proactive action rather than reactive response.
Vintage analysis in NBFC lending is the practice of grouping loans by their origination period (the “vintage”) and tracking each cohort’s performance over time on a consistent timeline. Instead of asking “what is our NPA today?”, vintage analysis asks “how is each set of loans performing at the same age, and how do different cohorts compare to each other?”
This guide explains what vintage analysis is, how vintage curves are built, what they reveal about portfolio quality and underwriting effectiveness, and how they function as early warning signals for credit teams.
Why Aggregate NPA Ratios Are Not Enough
Imagine an NBFC with a stable Gross NPA ratio of 3.5% over four consecutive quarters. That stability looks reassuring. But behind that stable aggregate number, something else could be happening: the 2023 vintage (loans originated in 2023) is performing beautifully at 1.2% NPA at 18 months of age. But the 2024 vintage is already showing 4.8% NPA at the same 18-month mark. The newer loans are deteriorating significantly faster.
This divergence does not show in the aggregate NPA ratio because the well-performing 2023 vintage offsets the deteriorating 2024 vintage. By the time the 2024 deterioration drags the aggregate ratio up, the problem may be 12 months old and thousands of crores deep.
Vintage analysis would have flagged this divergence at month 9 or 12 of the 2024 vintage early enough to investigate and intervene before the damage accumulated.
What a Vintage Is and How Cohorts Are Defined
Vintage in lending is the time period during which a group of loans originated. Vintages can be defined by:
- Monthly vintage: all loans originated in January 2024, all loans in February 2024, etc. Provides the most granular view but can produce noisy curves for small-volume lenders.
- Quarterly vintage: all loans originated in Q1 2024, Q2 2024, etc. Smooths seasonal variation and provides cleaner curves while maintaining enough granularity to detect underwriting changes.
- Annual vintage: all loans originated in 2023, 2024, etc. Useful for broad portfolio trend analysis but too coarse to detect specific underwriting changes or policy shifts.
For most NBFCs with monthly disbursement volumes above Rs 50 crore, monthly or quarterly vintages provide sufficient granularity. For smaller NBFCs, quarterly vintages are more practical.
Building the Vintage Curve: A Step-by-Step Explanation
The vintage curve plots the cumulative NPA rate for a cohort against the cohort’s “months on book,” the age of the loans since origination.
Construction:
- Define the cohort: select all loans originated in, say, Q2 2024.
- Track NPA formation: for each subsequent month, record what percentage of the cohort’s original outstanding has become NPA.
- Cumulative NPA calculation: at month 3 of the cohort’s life, what percentage has gone NPA? At month 6? At month 12? At month 18? These data points form the curve.
- Plot against months on book: the x-axis is months on book (1 through 36+). The y-axis is cumulative NPA as a percentage of original outstanding. The curve rises over time as more loans deteriorate.
- Compare across vintages: overlay Q1 2024, Q2 2024, Q3 2024, and Q4 2024 curves on the same chart. At each month-on-book point, directly compare how different cohorts were performing at the same age.
The vintage curve shape for a healthy product: rises steeply in the first 6–12 months (early delinquencies from origination errors and borrower misrepresentation), then flattens as the remaining portfolio stabilises. The cumulative NPA at 24 months on book is a commonly used single-number benchmark for comparing vintages.
What Vintage Curves Reveal About Portfolio Quality
Vintage curve analysis reveals five types of information about portfolio quality:
- Absolute performance level: Is the vintage’s cumulative NPA at 24 months on book consistent with the target? A product designed for 3% cumulative NPA at 24 months performing at 5.5% signals underwriting or product risk that the absolute target did not account for.
- Vintage deterioration: Are newer vintages showing higher NPA at the same age than older vintages? If the Q1 2024 vintage showed 2.1% NPA at 12 months, but the Q3 2024 vintage already shows 3.8% NPA at 12 months, something changed in underwriting, product, or borrower quality between Q1 and Q3.
- Vintage improvement: Are newer vintages performing better than older ones? If the 2025 vintages are consistently outperforming 2024 vintages at the same age, an underwriting improvement (new bureau integration, AA adoption, improved income verification) is working.
- Seasonality effects: Q4 vintages (loans originated in the year’s last quarter) may systematically perform differently from Q1 vintages, potentially because origination decisions are influenced by year-end volume targets. Vintage analysis isolates this seasonal effect.
- Macro effects: a cohort that originated during a period of economic stress (pandemic, sector downturn) shows different vintage curves from cohorts that originated in benign conditions. Vintage analysis separates origination-period risk from portfolio management.
How Underwriting Changes Show Up in Vintage Analysis
Every time an NBFC changes its underwriting policy, raises the minimum CIBIL floor, reduces the maximum FOIR, adopts Account Aggregator, or adds a new bureau check, the impact shows up in the performance of the first vintage originated under the new policy compared to the last vintage under the old policy.
This makes vintage analysis the most direct feedback mechanism for underwriting quality. If an NBFC raised its CIBIL floor from 680 to 720 in March 2024, the April 2024-onward vintages should show lower NPA curves than the January–March 2024 vintages, assuming the floor change was the primary driver of NPA formation.
If the vintages after the policy change do not show improvement, either the change was insufficient, the problem was not the CIBIL floor, or other factors are masking the improvement. Vintage analysis provides the evidence to distinguish between these interpretations.
Vintage Analysis as an Early Warning System
Early warning through vintage analysis uses the shape of an early-stage vintage curve to project future NPA levels before they materialise in the portfolio.
Established portfolios develop vintage performance benchmarks: the typical NPA rate at month 3, month 6, month 9, month 12. A new vintage whose month-6 NPA is already 40% above the month-6 benchmark is on a trajectory to significantly exceed the benchmark at month 24. The early divergence is the signal.
Early warning triggers based on vintage curves:
- Month 3 NPA above 2× benchmark indicates a large number of first-payment defaults, suggesting significant origination quality issues or fraud in the cohort.
- Month 6 NPA above 1.5× benchmark: indicates early stress forming across the cohort, likely from income verification errors or borrower over-extension.
- Month 12 NPA diverging from benchmark by more than 1 percentage point: a material and sustained underperformance that signals the cohort will significantly exceed target NPA at maturity.
Early warning triggers prompt specific actions: collections team review of the cohort, credit policy review to identify the origination characteristic that correlates with the poor performance, and sometimes temporary tightening of credit criteria while the cause is investigated.
Key Takeaways
- Vintage analysis in NBFC lending groups loans by origination period and tracks each cohort’s cumulative NPA on a months-on-book timeline, revealing what aggregate NPA ratios cannot: which cohorts are underperforming, when the deterioration began, and what may have caused it.
- Vintage curves plot cumulative NPA (y-axis) against months on book (x-axis) for each cohort, allowing direct comparison of different cohorts at the same age.
- Key insights from vintage analysis: absolute performance level, vintage deterioration or improvement, seasonality effects, and macro origination-period effects.
- Underwriting policy changes show up in vintage performance gaps between pre-change and post-change cohorts, making vintage analysis the most direct feedback loop for credit policy effectiveness.
- Early warning: month-3 NPA above 2× benchmark signals first-payment default issues; month-6 above 1.5× benchmark signals early cohort stress; month-12 divergence above 1 percentage point from benchmark signals material underperformance.
Frequently Asked Questions
What is vintage analysis in loan portfolio management? Vintage analysis groups loans by their origination period (monthly, quarterly, or annually) and tracks each group’s credit performance on a consistent age-based timeline. It plots cumulative NPA as a percentage of original outstanding against the number of months since origination (months on book). This allows direct comparison of how different cohorts are performing at the same age revealing deteriorating trends, improving trends, and the impact of underwriting or policy changes that aggregate NPA ratios cannot show.
How is a vintage curve built? A vintage curve takes all loans originated in a specific period (the “vintage”), then tracks what percentage of that cohort’s original outstanding has become NPA at each subsequent month. Month 3: what percentage of the Q1 2024 cohort is NPA? Month 6? Month 12? These data points are plotted with months on the book on the x-axis and cumulative NPA percentage on the y-axis. Multiple vintage curves can be plotted on the same chart to compare cohort performance visually.
What does a worsening vintage curve indicate for an NBFC? A worsening vintage curve where newer cohorts show higher NPA rates at the same months-on-book as older cohorts indicates that credit quality is deteriorating. The cause may be: relaxation of underwriting standards, a higher-risk borrower mix, a macro deterioration in the borrower segment, fraud in a specific channel or geography, or a combination. Vintage analysis reveals the timing of the deterioration; subsequent root cause analysis identifies the specific driver.
At what point on the vintage curve should an NBFC take action? Most NBFCs set early warning triggers based on benchmark vintage curves. Common thresholds: month-3 NPA above 2× the product benchmark (signals first-payment default issues requiring immediate cohort review), month-6 NPA above 1.5× benchmark (signals early systematic stress), and month-12 NPA more than 1 percentage point above benchmark (signals material cohort underperformance). When these triggers fire, collections review of the cohort, credit policy review, and potential temporary underwriting tightening are initiated.
How often should vintage analysis be conducted? Vintage analysis should be conducted monthly as part of the portfolio management cadence. Monthly monitoring allows early warning triggers to fire at the earliest possible point. A quarterly vintage analysis review including trend analysis across the past 4–6 quarters of vintage performance is typically presented to the Board-level Risk Management Committee as part of the credit quality reporting pack.
Conclusion
Vintage analysis in NBFC lending is the analytical practice that converts portfolio management from reactive to proactive. Aggregate NPA ratios tell you what the portfolio looks like today. Vintage curves tell you what it is becoming and give you the lead time to act before it gets there.
Make vintage analysis a regular, automated part of your portfolio reporting infrastructure. The NBFC that identifies a deteriorating vintage at month 6 and investigates the cause can limit the damage. The one that sees it at month 18 through the aggregate NPA ratio is already managing consequences, not preventing them.
Verify smarter. Lend safer with Fineye.
August 22, 2026
9 min read
