Loan Collections Benchmarks India 2026: Contact, PTP, and Recovery Rates
Loan collections benchmarks in India for 2026 combine two layers: macro asset-quality data — banks' gross non-performing assets (NPAs) at a multi-decade low of 2.3% in 2024-25 ([RBI Financial Stability Report, 2025](https://www.business-standard.com/finance/news/rbi-npa-reduction-driven-by-write-offs-in-last-five-years-fsr-125063001214_1.html)) — and operational metrics such as contact rate, right-party contact, promise-to-pay, and roll rates that lenders track daily.
By the YuVerse Editorial Team · Published 2026-07-30
This is an explainer, not legal, financial, or regulatory advice. Public benchmarks for internal collections metrics are thin, so we cite official macro data where it exists and describe operational rates as industry ranges — never as invented precision.
Why Do Collections Benchmarks Matter in 2026?
Collections sits between a lender's balance sheet and its borrowers. When repayment slips, the health of the loan book shows up first in operational dashboards — who picks up the phone, who promises to pay, who actually pays — and only later in reported NPAs.
The macro backdrop is reassuring at the top line but uneven underneath. Banks' gross NPA and net NPA ratios fell to multi-decadal lows of 2.3% and 0.5% respectively in 2024-25, per the Reserve Bank of India's (RBI) Financial Stability Report. But the RBI also noted that write-offs were a major driver of that decline — the write-offs-to-gross-NPA ratio rose to 31.8% in FY25 (RBI FSR via Business Standard, 2025).
Underneath the headline, unsecured retail lending is where collections teams feel the most pressure — the theme running through our AI for banking collections India playbook.
What Are the Core Collections Metrics? (Glossary)
Before comparing numbers, teams need shared definitions. Here is the working glossary most Indian lenders use.
Metric | What it measures |
|---|---|
Contact rate | Share of accounts where any human is reached on a dial or campaign |
Right-party contact (RPC) | Share of contacts where the actual borrower (not a relative or wrong number) is reached |
Promise-to-pay (PTP) | Share of right-party contacts where the borrower commits to a payment amount and date |
Kept-PTP | Share of promises that convert into an actual payment by the due date |
Roll rate | Share of accounts that move from one delinquency bucket to a worse one |
Cure / back-flow rate | Share of accounts that move back to a healthier bucket after payment |
Recovery rate | Value recovered as a share of the outstanding amount in a bucket or pool |
Days Past Due (DPD) buckets are the spine of collections. Accounts are grouped by how overdue they are — typically 1-30, 31-60, 61-90, and 90+ DPD — and each bucket has its own strategy, script, and expected outcome.
What Do the DPD Buckets Tell You?
Behaviour changes sharply as accounts age. Early buckets (1-30 DPD) are mostly forgetfulness or short cash-flow gaps, and respond to reminders. Later buckets (90+ DPD, the edge of an NPA) skew toward genuine stress or unwillingness to pay.
The RBI's own data shows this ageing risk. It flagged that first defaults are mostly seen in unsecured advances, and that slippage from Special Mention Account-2 into NPAs is rising while upgradations decline (RBI, via Business Standard, 2024).
DPD bucket | Typical borrower behaviour | Collections priority |
|---|---|---|
1-30 DPD | Missed reminder, short cash gap | High-volume reminders, self-cure |
31-60 DPD | Emerging stress, partial payers | PTP capture, restructuring options |
61-90 DPD | Persistent stress or dispute | Field + tele-calling, settlement talks |
90+ DPD (NPA edge) | Serious stress or unwillingness | Legal notices, recovery, write-off review |
For a step-by-step view of how lenders sequence these buckets, see our guide on how Indian banks use voice AI for outbound collections.
What Are Realistic Operational Benchmarks in India?
Public, audited benchmarks for contact rate, RPC, PTP and kept-PTP do not exist the way NPA data does — every lender defines and measures them slightly differently. So the honest answer is a set of directional ranges, not exact figures.
- Contact and RPC vary widely by product, DPD bucket and data quality. Early-bucket digital and voice campaigns reach far more borrowers than deep-bucket field visits; RPC is always lower than raw contact rate because of wrong numbers and third-party pickups.
- PTP and kept-PTP are the real signal. A promise is easy to capture; a kept promise is what protects the book. Kept-PTP is materially lower than PTP, and the gap widens in later buckets.
- Roll rates rise as stress builds. The direction — not a single number — is what teams manage against week to week.
Rather than publish invented precision, mature teams benchmark themselves against their own trailing performance and segment-level cohorts. That is the defensible way to read collections data in India in 2026.
Where Is the Stress Concentrated?
Two segments deserve a closer look because public data is unusually clear.
Unsecured retail. The RBI reported that unsecured retail lending, while moderating to 25% of retail loans and 8.3% of gross advances, had a weaker gross NPA ratio of 1.8% versus 1.2% for the overall retail portfolio in March 2025, with private banks driving fresh slippages (RBI FSR, 2025). Small-ticket personal loans are especially fragile: about 11% of borrowers taking personal loans under ₹50,000 already had an overdue personal loan, and over 60% had availed more than three loans (RBI, 2024).
Microfinance. Asset quality deteriorated sharply in FY25. Portfolio at Risk (PAR) 31-180 rose to 6.2% at the end of Q4FY25 from 2% a year earlier, NPAs climbed to nearly ₹55,000 crore — about 14.8% of gross loans — and the near-NPA PAR 91-180 bucket worsened to 3.5% from 0.9% (MFIN via Business Standard, 2025).
Segment | Key FY25 metric | Source |
|---|---|---|
Overall banks (SCBs) | GNPA 2.3%, NNPA 0.5% | RBI FSR, 2025 |
Unsecured retail | GNPA 1.8% vs 1.2% overall retail | RBI FSR, 2025 |
Microfinance | NPAs ~₹55,000 crore (14.8% of loans) | MFIN, 2025 |
Microfinance | PAR 1-90 at 4.22% (Mar 2025) | MFIN, 2025 |
How AI Helps
Collections economics turn on contact and follow-through at scale — exactly where voice AI fits. YuVoice runs multilingual, compliant outbound reminders and early-bucket campaigns, capturing promise-to-pay commitments and routing only the hard cases to human agents. Across the YuVerse platform, voice AI handles 2.5 crore calls a month, which lets lenders dial every early-DPD account consistently rather than sampling. Consistent contact in the 1-30 DPD window is the single biggest lever on kept-PTP and roll rates — automating it frees human collectors for negotiation-heavy deep buckets. See how teams automate payment reminders with AI voice calls.
FAQ
What is a good right-party contact rate for Indian lenders? There is no single audited benchmark. RPC depends heavily on data quality, DPD bucket and channel. The practical approach is to track RPC against your own trailing cohorts and improve number hygiene, call timing, and multilingual scripting rather than chase a published figure.
How is promise-to-pay different from kept-PTP? Promise-to-pay is a borrower's verbal commitment to pay a specific amount by a date. Kept-PTP measures how many of those promises actually convert into payment. Kept-PTP is always lower, and the gap is the real health indicator for a collections book.
Are India's NPAs really at a record low? Yes at the headline level — the RBI reported gross and net NPA ratios at multi-decadal lows of 2.3% and 0.5% in FY25. But the RBI also cautioned that write-offs, which rose to 31.8% of gross NPAs, were a major driver of the decline (RBI FSR, 2025).
Why is microfinance collections under pressure in 2026? MFIN data shows PAR 31-180 rising to 6.2% and NPAs near ₹55,000 crore by March 2025, driven by overleveraging, heatwaves and state-level recovery restrictions (MFIN, 2025). Contact and PTP discipline matter more than ever in this segment.
What are roll rates and why track them? Roll rate is the share of accounts moving from one DPD bucket to a worse one — for example, 31-60 into 61-90. Rising roll rates signal deteriorating collections and future NPAs, so teams manage them bucket by bucket.
Can AI replace human collectors? No. AI handles high-volume, low-complexity early-bucket outreach and reminders at scale; human agents handle negotiation, disputes and settlements in deep buckets. The two work together — see AI voice agents for loan collections use cases.
Conclusion
India's collections benchmarks for 2026 are a two-layer story: strong headline asset quality masking pockets of stress in unsecured retail and microfinance. The operational metrics — contact, RPC, PTP, kept-PTP and roll rates — are where lenders win or lose, and where consistency at scale beats one-off effort.
Benchmark your collections book against the metrics that matter. Talk to the YuVerse team to see how AI voice agents lift early-bucket contact and promise-to-pay.
References
- RBI Financial Stability Report (write-offs and NPA data), 2025 — https://www.business-standard.com/finance/news/rbi-npa-reduction-driven-by-write-offs-in-last-five-years-fsr-125063001214_1.html
- RBI on small personal-loan slippages, 2024 — https://www.business-standard.com/economy/news/rbi-flags-risk-to-secured-loans-from-slippages-in-smaller-personal-loans-124123000827_1.html
- MFIN microfinance portfolio quality (PAR and NPAs), 2025 — https://www.business-standard.com/finance/news/microfinance-portfolio-quality-dips-par-npas-rise-fy25-mfin-125061101166_1.html
