This paper outlines the development of a practical approach to simulating a credit loss distribution function and to implementing a stress test exercise focusing on the entire Spanish mortgage portfolio. Specifically, we determine, via regression model, the main factors that explain why households fail to meet their mortgage payment commitments. This allows us to assign individual borrowers’ PDs and to estimate a rating system for the mortgage portfolio. Then, we simulate the empirical distribution function of mortgage loss rates using a Monte-Carlo resampling method, and compare the loss rates from this function with those provided by the Basel II IRB formulas. Finally, we assess, by running a stress exercise, the ability of banks to withstand certain adverse situations. The main result from this exercise is that, in general terms, Basel II IRB regulatory loss coverage offers fairly adequate protection for banks.

We use European Banking Authority (EBA) 2014 stress test data to study the use of the internal ratings based approach (IRB) and the risk weights of European banks. A simple inspection of data at country level reveals significant differences in the use of the IRB approach by banks and in risk weighted asset (RWA) densities. Empirical analysis of these differences is conducted at both the level of bank group and country of exposure. A clearly negative relation is found between use of the IRB approach and RWA densities, even after controlling for portfolio and bank characteristics. The jurisdiction in which a credit exposure is located affects both the use of IRB and risk density. Portfolio variables (default rate, weight of corporate exposures), and bank characteristics (book equity ratio, liquidity measures) are also significantly related to risk weights. The analysis of the dataset also provides some evidence of a relation between RWA density and IRB utilization, and portfolio composition, bank size, and market risk measures (stock return volatility, CDS spreads).