What NPL means

Granular portfolios gather thousands of low-value contracts — cards, personal loans, payroll loans, vehicles. Analysis is statistical: it works with behaviour curves by vintage and product, not with individual loans.

Corporate exposures concentrate material value in few debtors. Here analysis is case by case: collateral, subordination, business viability and reorganisation scenarios.

These pricing logics are incompatible. Applying the statistical model to a corporate exposure, or analysing a hundred thousand contracts individually, are symmetrical errors.

Aging: the older the delinquency, the lower the probability of recovery.

Product and origination: collateralised credit recovers differently from unsecured credit.

The portfolio's own recovery history, where available — the most informative data point and the most frequently missing.

Documentary soundness: enforceability depends on the integrity of the instrument and the assignment chain.

The capacity of the servicer that will take over collection.

Two identical portfolios under different servicers produce materially different recovery curves. The collection workflow, systems, data quality and operational discipline account for much of the outcome.

That is why servicer assessment forms part of portfolio diligence rather than a later stage: a price is only defensible once you know who will operate the asset.

WARR — weighted average recovery rate for the portfolio.

Recovery curve by month post-assignment, typically projected from M+1 to M+36, or to M+60 for long portfolios.

Collection cost over recovered revenue, which defines the effective margin.

Average ticket, average aging and the share of the portfolio with collateral.

Assignment of retail portfolios involves personal data processing at scale. Legal basis, purpose, security in transfer and the servicer's compliance must be addressed during structuring — not after closing.