How the credit risk index is calculated
Components and weights
| Component | Value used | Weight |
|---|---|---|
| NPL ratio | Latest banking-system NPL ratio | 25% |
| Change in NPL ratio | Change from the same period a year earlier (points) | 20% |
| Credit growth | Distance of year-on-year credit growth from the median of the eight markets (fast growth and contraction both add risk) | 20% |
| Household debt | Household debt to GDP, where published | 15% |
| Policy rate direction | Last three months: raised +1, unchanged 0, cut −1 | 10% |
| Regulatory tightening | Last 90 days: rules that tighten minus rules that ease (drafts count half) | 10% |
Normalisation
For each component, the values of the markets with data are standardised (mean 0, standard deviation 1), capped at ±2, and turned into a 0–100 score as 50 + 25 × z. The market score is the weighted average of its component scores. A higher score means harder conditions for lenders.
Missing values
A component with no figure from an allowed source is left out and the other weights are scaled up in proportion. A market with fewer than four of the six components gets no score or rank that month. Credit growth reported only as year-to-date is not used, because it cannot be compared with year-on-year figures.
Sources
Central banks, financial regulators, statistics offices and finance ministries, and international organisations (IMF, BIS, World Bank, ADB, AMRO). No media reports or private data. Each component shows its source and release date for members. Regulatory changes are counted from T&S reports, each based on the authority’s own document.
Updates
Updated once a month. Each edition uses the latest figures released by then, so markets with less frequent data move less often. Changes to the method are recorded on this page with the date.