Method
Every figure on this site comes from one of the definitions below. Where the Reserve Bank prescribes the answer we cite the paragraph; where it does not, we say plainly that the choice is ours and what we chose.
This is the part we publish. The queries and the calculation code stay on our side, but a definition you cannot inspect is a number you cannot trust — so the definitions are here in full, including the ones where reasonable people report differently.
These are the metrics where two honest lenders report different numbers for the same book. Every one of them is toggleable on the dashboards, so you can watch your figure move.
Share of the book classified non-performing.
Share of accounts classified non-performing.
GNPA where one bad loan makes all that borrower's loans non-performing.
Of what fell due this month how much was collected this month.
All cash collected this month against what fell due this month.
All cash collected against this month's dues plus arrears carried in.
Gross collections plus recoveries on written-off loans.
Of the instalments due how many were paid in full.
Share of the book owed by borrowers more than 30 days late.
Share of the book owed by borrowers more than 60 days late.
Share of the book owed by borrowers more than 90 days late.
Accounts 31 to 60 days late but not yet non-performing.
Accounts 61 to 90 days late but not yet non-performing.
Non-performing for up to twelve months or restructured within the last year.
Non-performing for more than twelve months.
Identified as unrecoverable but still on the books.
The same delinquency counted from the day after the due date.
The full register, read from the same file the summary layer is written against — so this page cannot drift from what the site actually computes.
| Metric | What it means | Basis | Reference |
|---|---|---|---|
Principal outstanding Sum of closing principal across live accounts. | What borrowers still owe in principal on live loans. | Derived | Engine C-DV-04 movement identity |
Disbursals Sum of amounts disbursed in the month. | Principal lent in the month. | Derived | — |
Live accounts Count of accounts. | Number of loan accounts with principal still outstanding. | Derived | — |
PAR 30+ Principal of accounts over 30 DPD divided by principal outstanding. | Share of the book owed by borrowers more than 30 days late. | Our choice | Industry convention; not an RBI term |
PAR 60+ Principal over 60 DPD divided by principal outstanding. | Share of the book owed by borrowers more than 60 days late. | Our choice | Industry convention |
PAR 90+ Principal over 90 DPD divided by principal outstanding. | Share of the book owed by borrowers more than 90 days late. | Our choice | Industry convention. Close to GNPA but not identical: PAR counts days late; GNPA counts classification |
GNPA by value Principal of non-standard accounts divided by principal outstanding. | Share of the book classified non-performing. | RBI | IRACP 2025 paras 19 and 21 |
GNPA by number Non-standard accounts divided by live accounts. Usually higher than GNPA by value because bad loans are smaller. | Share of accounts classified non-performing. | RBI | IRACP 2025 paras 19 and 21 |
GNPA borrower-wise Borrower-wise classification rather than facility-wise. | GNPA where one bad loan makes all that borrower's loans non-performing. | RBI | IRACP 2025 para 23: borrower-wise and not facility-wise |
NNPA Non-performing principal less provision held divided by principal outstanding less provision held. | GNPA after subtracting the provision already held against it. | RBI | IRACP 2025 para 32 |
Provision coverage Provision divided by non-performing principal. | Provision held as a share of the non-performing book. | RBI | IRACP 2025 para 32 |
Days past due Counted from the due date itself as day one. | Days since the oldest unpaid instalment fell due. | RBI | IRACP 2025 para 11(2) and Illustration I: SMA-1 at 30 Apr, SMA-2 at 30 May, NPA at 29 Jun |
Days past due (day-0 counting) One day fewer than day-1 counting. Shown only to size the gap against systems that count this way. | The same delinquency counted from the day after the due date. | Our choice | Not RBI. Shown because some systems report it |
Collection efficiency (current month) Collections against this month's instalments divided by this month's demand. | Of what fell due this month how much was collected this month. | Our choice | No single RBI definition. The strictest of the five |
Collection efficiency (gross) Total collections divided by current demand. Can exceed 100% when arrears are recovered. | All cash collected this month against what fell due this month. | Our choice | Most commonly quoted. The reason a CE figure can pass 100% |
Collection efficiency (demand plus arrears) Total collections divided by current demand plus opening arrears. The most conservative denominator. | All cash collected against this month's dues plus arrears carried in. | Our choice | — |
Collection efficiency (including recoveries) Adds post-write-off recovery to the numerator. Answers to no demand so it inflates the ratio. | Gross collections plus recoveries on written-off loans. | Our choice | — |
Billing efficiency Counted on instalments rather than money so a part payment does not count. | Of the instalments due how many were paid in full. | Our choice | — |
Demand Excludes schedule rows superseded by a restructure, and instalments falling due after the loan closed. | Instalments that fell due in the month and were actually owed. | Derived | Counting every schedule row overstates demand by about 5 points of CE |
Interest accrued Sum of interest on instalments falling due in the month. | Interest charged for the month on the contract. | Derived | — |
Interest recognised Accrual while performing; cash basis once non-performing. | Interest taken to income. On a non-performing loan only what was actually received. | RBI | IRACP 2025 para 38: income on an NPA is recognised only on realisation |
Interest reversed Reversed in the month the loan becomes non-performing. | Income previously recognised but unrealised reversed when a loan turns non-performing. | RBI | IRACP 2025 para 38 |
Yield Interest recognised divided by average of opening and closing principal. | Interest recognised as a share of average principal outstanding. | Derived | Differs from the contracted rate because of para 38 reversals |
Write-offs Sum of amounts written off in the month. | Principal removed from the book as unrecoverable. | Our choice | RBI mandates provisioning; write-off timing is board policy |
Post-write-off recovery Recoveries divided by cumulative write-offs. | Cash recovered on loans already written off as a share of what was written off. | Derived | — |
SMA-1 Pre-NPA warning stage. An account already non-performing carries no SMA label. | Accounts 31 to 60 days late but not yet non-performing. | RBI | IRACP 2025 para 19 |
SMA-2 | Accounts 61 to 90 days late but not yet non-performing. | RBI | IRACP 2025 para 19 |
Sub-standard | Non-performing for up to twelve months or restructured within the last year. | RBI | IRACP 2025 para 52: 52(2) makes a restructured asset sub-standard until a year of satisfactory performance |
Doubtful | Non-performing for more than twelve months. | RBI | IRACP 2025 para 53 |
Loss Provided for in full. | Identified as unrecoverable but still on the books. | RBI | IRACP 2025 paras 21 and 32 |
The figures on this site are computed by an engine that is tested rather than trusted. Every rule it applies is a row in a register with its source and status, not a number buried in code. The engine is covered by a suite of hand-worked cases, and — more usefully — by a harness that deliberately plants known errors and confirms the tests catch every one. A test suite that passes proves very little; one that provably fails when the code is wrong proves something.
The same discipline produces the findings we hand a client: where two of your own reports disagree, we can show which definition each is on, and what the number becomes under the other.