the web's biggest credit risk modeling resource.

Credit Jobs

Home Glossary Links FAQ / About Site Guide Search


Submit Your Paper

In Rememberance: World Trade Center (WTC)

doi> search: A or B

Export citation to:
- Text (plain)
- BibTeX

Assessing Credit Loss Distributions: Bayesian Multi-Period Model vs. Basel II Model

by Leonid V. Philosophov of Moscow Committee of Bankruptcy Affairs

August 9, 2004

Abstract: As development of the New Basel Capital Accord (Basel II) approaches to its final, the problem of validation and calibration of the Basel II models attracts increasing interest of scientists and practitioners. This is evidenced by the Basel Committee press releases, recent past and forthcoming conference programs, scientific publication on the problem.

The principal attention of the Basel II is concentrated on credit risk where validation problems seem to be solvable, because credit risks of individual borrowers and portfolios are accessible for precise calculations.

The current paper assesses validity of credit loss distributions, calculated by means of the Basel II model. The assessment method consists in parallel calculations the same distributions by means of exact probabilistic formulae. The exact calculation scheme is realized within Bayesian multi-period (BMP) model and Portfolio BMP model.

We found that Basel II model ensures correct assessments of credit loss distributions for medium and large portfolios of thirty or more one-period credits (credits without intermediate interest etc. payments). For small portfolios and individual one-period credits assessments of Basel II model are rough or extremely rough. Basel II model undervalues credit risk of portfolios of multi-period credits (credits with active period covering several years and intermediate payments).

The study provides an alternative look at some principal Basel II concepts, like Probability of Default (PD) and Correlation. This can serve to more correct understanding and flexible use of those concepts in assessing credit risks.

Exact models are more complex that Base ll model, but perfectly available for quick computer calculations. They can be used in bank practice ensuring flexible assessments of credit risk.

JEL Classification: C11, C13, C51, E58.

Keywords: Credit risk, Credit loss distribution, Basel II, Bayesian model.

Books Referenced in this paper:  (what is this?)

Download paper (405K PDF) 25 pages