Institutional Repository of Vilnius University
Koreliuotų juridinių asmenų įsipareigojimų nevykdymo tikimybės įverčių pasikliovimo intervalų analizė /
Abstract
dc:descriptionIn the modern financial environment, credit risk is one of the most important risks faced by financial institutions, and one of its key assessment components is the probability of default (PD), along with the uncertainty associated with its estimate. In practice, this uncertainty is often assessed using traditional confidence intervals, such as the Wald, Clopper-Pearson, Wilson and Agresti-Coull intervals. However, these methods rely on the assumption of independent events. Since borrowers in real credit portfolios may be affected by common macroeconomic factors, this assumption is not always justified. This study analyses the properties of confidence intervals for probability of default estimates of legal entities under independent and correlated default events, using synthetic data ge nerated by the Monte Carlo method with predefined PD values, different portfolio sizes and asset correlation levels, while the dependence between borrowers is modelled using a onefactor Vasicek model. For each scenario, the main properties of the confidence intervals are evaluated: empirical coverage probability and interval width. The results show that, under independent events, most classical intervals perform relatively stably. However, under correlated events, their coverage probability decreases significantly, while the interval width does not increase accordingly. The study also shows that the Wald interval is particularly sensitive to small PD values and scenarios with a low expected number of default events. In addition, the study evaluates the possibility of applying an effective sample size correction, which allows the effect of correlated events on the properties of confidence intervals to be partially compensated. Finally, possible directions for further research are discussed.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Lozenko, Makar,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD308334696&prefLang=en_US