27.03.2015 Views

BASEL II: PROBLEMS AND USAGE

BASEL II: PROBLEMS AND USAGE

BASEL II: PROBLEMS AND USAGE

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Small and medium-sized universal banks prefer to use the less sophisticated<br />

Standard Approach.<br />

Hence, despite the fact that Quantitative Impact Studies (QIS), which<br />

were conducted to assess the impact of Basel <strong>II</strong> on the banks’ capital requirements,<br />

revealed that the use of the advanced methods would be especially<br />

beneficial for small and medium-sized banks, 25 these banks prefer to<br />

use the Standard Approach. It is worth mentioning that there is a possibility<br />

of partial use of the IRBA, according to which the banks can implement the<br />

rating systems on a step-by-step basis within a period of five years. Actually,<br />

the average implementation period in case of the partial use constitutes<br />

three years.<br />

Although the number of rating systems submitted for approval ranges<br />

from one to 50, the most common methods are expert systems, simulation<br />

models, and credit scoring systems. 26 The latter use quantitative and qualitative<br />

data in order to assign a score to a borrower, which reflects his creditworthiness.<br />

The most popular statistical tools employed are discriminate<br />

analysis and probability models. 27 Scoring systems are primarily used in<br />

retail business and for small and medium-sized enterprises. Expert systems<br />

are applied to evaluate the credit risk of large corporate borrowers. Finally,<br />

simulation methods are employed for specialized lending and project finance.<br />

In practice, a mixture of all three methods is also used. 28<br />

5.2. Model Validation Techniques<br />

Whichever of the aforementioned internal ratings-based methods is<br />

applied to model the credit risk, it has to be of high quality to ensure the<br />

correct reflection of the bank’s credit risk exposure. Various validation<br />

techniques were developed to evaluate the quality of internal rating models.<br />

Basel <strong>II</strong> devotes special attention to the validation issue; banks applying for<br />

IRBA implementation have to convince the supervisor in the appropriateness<br />

of their models. One way of validating the rating model is to measure<br />

its discriminative power, i.e. the ability to distinguish between defaulters<br />

and non-defaulters. The most popular validation techniques include the<br />

Cumulative Accuracy Profile (CAP) and the Receiver Operating Characteristic<br />

(ROC). 29<br />

25 See Deutsche Bundesbank (2006).<br />

26 See Deutsche Bundesbank (2009).<br />

27 These models include the linear probability model, the logit model, and the probit<br />

model.<br />

28 See Deutsche Bundesbank (2009).<br />

29 This subsection is based on Beinert/Reichling/Vogt (2007).<br />

30

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!