Non-performing loan (NPL)

A non-performing loan (NPL) is a loan on which the borrower has failed to make scheduled payments for a defined period — conventionally 90 days or more — or on which full repayment is considered unlikely regardless of arrears status.

The definition rests on two independent tests, and a loan meeting either is non-performing:

1. The quantitative test — payments of principal or interest are past due by 90 days or more. The threshold is a widely used convention, embedded in Basel guidance and most national supervisory frameworks, though some regimes apply different thresholds to particular exposure types.
2. The qualitative test — "unlikely to pay" (UTP) — the lender assesses that the borrower is unlikely to repay in full without recourse to collateral realisation, irrespective of days past due. Triggers include borrower insolvency, business closure, distressed restructuring, absconding, or the sale of the exposure at a material credit-related loss.

The second test matters more than it is usually given credit for. A loan can be non-performing with zero days past due — a borrower servicing debt from new borrowing rather than from operations is unlikely to pay, whatever the arrears schedule shows.

The NPL Ratio

NPL ratio = Non-performing loans / Gross loan portfolio × 100

The numerator is the full outstanding balance of loans classified as non-performing, not merely the overdue instalments. This convention is essential and frequently misapplied.

Worked example

A lender has a gross loan portfolio of 10,000,000. Loans with at least one instalment more than 90 days overdue have a combined outstanding balance of 700,000, of which the actually overdue instalments total 180,000.

• Full outstanding balance (correct convention): 700,000 / 10,000,000 = 7.0%
• Overdue instalments only (incorrect): 180,000 / 10,000,000 = 1.8%

The difference is nearly fourfold on identical data. Counting only overdue instalments dramatically understates the problem, because it ignores the fact that the remaining balance of a defaulted loan is also at risk. Any NPL or PAR figure should be confirmed as full-balance before it is relied upon.

Net NPL ratio

Net NPL ratio = (NPL − provisions held against NPL) / (Gross loan portfolio − provisions) × 100

The net figure shows unprovided exposure — the portion of non-performing loans not yet absorbed by the loss allowance, and therefore still capable of hitting future earnings and capital.

NPL coverage ratio

NPL coverage ratio = Loan loss allowance / Non-performing loans × 100

The standard adequacy check. Low coverage against a high NPL ratio indicates provisioning is lagging the deterioration already visible on the books.

NPL vs PAR: The Distinction

These two metrics are close relatives and are routinely conflated, particularly in microfinance where PAR is the dominant convention and NPL the regulatory one.

• Basis: PAR>90 relies purely on days past due; NPL relies on days past due or an unlikely-to-pay judgement.
• Origin: PAR is a microfinance reporting convention; NPL originates from banking supervision and accounting classification.
• Numerator: PAR uses the full outstanding balance of loans overdue beyond the threshold; NPL uses the full outstanding balance of loans classified non-performing.
• Restructured loans: PAR exposures are generally re-aged on the new schedule; NPL exposures often remain non-performing through a probation period.
• Judgement involved: None for PAR; qualitative UTP assessment required for NPL.
• Typical relationship: PAR is typically lower; NPL is typically higher or equal.

In practice PAR>90 and the NPL ratio converge closely for a simple portfolio with no restructuring and no UTP cases. They diverge where an institution restructures actively — because rescheduling resets the days-past-due clock that PAR depends on, while most NPL frameworks keep the exposure classified as non-performing until a probation period has been served.

A large and growing gap between PAR>90 and the NPL ratio is itself a signal, usually indicating restructuring volume that PAR is not capturing.

Loan Classification Bands

Most supervisory frameworks require loans to be classified into bands, with prescribed minimum provisioning attached. Terminology and thresholds vary by jurisdiction, but the structure is broadly consistent:

• Standard / Pass: Current, or minimal arrears (Performing)
• Special mention / Watch: Roughly 30–89 days past due, or early weakness signals (Performing but deteriorating)
• Substandard: Roughly 90–179 days past due (Non-performing)
• Doubtful: Roughly 180–359 days past due (Non-performing)
• Loss: Roughly 360+ days past due, or deemed uncollectible (Non-performing)

Illustrative. Thresholds, band names and prescribed provisioning rates are set by each regulator and must be confirmed locally.

Two features of these frameworks are worth noting:

• Whole-borrower classification. Many regimes require that if one facility of a borrower is non-performing, all facilities of that borrower are classified non-performing — sometimes extending to related parties. A single delinquent exposure can therefore reclassify a much larger book.
• Collateral treatment. Some frameworks permit the value of eligible collateral to be deducted before applying prescribed provisioning rates. This affects the provision, not usually the classification itself.

NPL, IFRS 9 Stage 3, and Default

Three overlapping but non-identical concepts:

• Non-performing is a supervisory classification, defined by the regulator or by the applicable reporting framework.
• Stage 3 under IFRS 9 means credit-impaired: objective evidence of impairment exists, lifetime expected credit loss applies, and interest is recognised on the net rather than gross carrying amount.
• Default is defined by the institution itself under IFRS 9 for staging and PD estimation, subject to a rebuttable presumption that default occurs no later than 90 days past due, and must be consistent with internal credit risk management practice.

In a well-designed framework these three populations should broadly coincide, and institutions are generally expected to explain material divergence. They are nonetheless distinct definitions with distinct owners, and the reconciliation between them is a routine audit and supervisory question.

Cure and Reclassification

A non-performing loan does not automatically return to performing status when a payment arrives. Standard practice requires a probation period — a defined span of consistent, on-schedule payments — before reclassification, precisely to prevent a single catch-up payment from resetting the status of a fundamentally impaired exposure.

Forborne exposures — those granted concessions for credit reasons — typically carry longer probation requirements, and are usually flagged as forborne for a period even after returning to performing status. Restructuring is a concession, not a cure, and treating it as one is the mechanism by which NPL ratios are most commonly understated.

What Moves the NPL Ratio

Because it is a ratio, both sides move it — and only one of the four drivers below reflects genuine improvement.

Numerator effects:

• New defaults increase it
• Cures decrease it (genuine improvement)
• Write-offs decrease it — removing the exposure entirely, without any recovery
• Restructuring can decrease it — where the concession moves the loan out of classification prematurely

Denominator effect:

• Rapid portfolio growth decreases it — new loans inflate the gross portfolio while defaults from those cohorts have not yet emerged, mechanically diluting the ratio

That last point deserves emphasis. A fast-growing lender will report a falling NPL ratio almost automatically, and the effect is strongest exactly when growth is outrunning underwriting discipline. The ratio is most flattering at the moment it is least informative. Vintage analysis — measuring default by origination cohort rather than by reporting period — is the standard correction.

Why NPLs Matter

• Earnings. Non-performing loans typically move to non-accrual status: interest stops being recognised as revenue, and under IFRS 9 stage 3 interest is calculated on the net carrying amount. The asset stops earning while continuing to consume funding.
• Capital. Provisions reduce profit and therefore retained earnings; unprovided NPLs represent future capital erosion.
• Funding cost. Lenders, depositors and rating agencies price NPL levels directly. Deterioration raises the cost of the institution's own borrowing.
• Management attention. Workout and recovery consume disproportionate senior time relative to the balances involved.
• Credit supply. At system level, high NPL stocks constrain lending capacity — a well-documented drag on credit growth in economies carrying legacy NPL burdens.
• Supervisory consequence. Breaching prescribed thresholds can trigger enhanced reporting, restrictions on dividends or growth, or directed remediation.

Managing Non-Performing Loans

• Early identification — the qualitative UTP test exists to catch deterioration before the 90-day threshold, and is only useful if applied actively rather than as a formality.
• Workout and restructuring, where the borrower has genuine capacity to service a revised schedule and the concession is properly classified as forbearance.
• Collateral realisation, where security exists and enforcement is economic.
• Write-off, once there is no reasonable expectation of recovery, with recovery efforts continuing off balance sheet.
• Portfolio sale — converting an uncertain long-tail claim into immediate discounted cash, with the trade-off of losing control over how a purchaser treats former clients.
• Root-cause analysis by vintage, product and branch. NPLs are an output of past underwriting; managing the stock without fixing the origination flow simply reproduces the problem.

Common Measurement Pitfalls

• Overdue instalments instead of full balance in numerator — understates the ratio several-fold.
• Gross versus net confusion — quoting a net NPL ratio against a gross benchmark, or vice versa.
• Ignoring restructured exposures — the single largest source of understated NPL ratios.
• Growth dilution — a falling ratio during rapid expansion often reflects the denominator, not credit quality.
• Write-off policy differences — an institution writing off at 180 days will report a structurally lower NPL ratio than one writing off at 360, with no difference in underlying performance. NPL ratios are not comparable between lenders without knowing both write-off policies.
• Inconsistent borrower-level aggregation — whether classification is applied per facility or per borrower materially changes the reported figure.

Frequently Asked Questions

Is a non-performing loan the same as a bad debt? No. Non-performing means repayment is significantly overdue or doubtful; the loan remains on the balance sheet and may still be cured or recovered. A bad debt written off has been derecognised as uncollectible.

What is a good NPL ratio? There is no universal benchmark. Acceptable levels differ sharply by product, borrower segment and pricing — an unsecured microloan book priced for the risk sustains a far higher ratio than a collateralised SME book. Comparisons are only meaningful between institutions with similar products and similar write-off policies.