Reporting and management system based on internal risk credit models according to Basel II capital agreement

Prema smjernicama koje propisuje sporazum Basel II, banke će se poticati na izgradnju sustava za upravljanje i izvješćivanje koji će na temelju internih modela banci pomoći u procjeni kreditnih, tržišnih i operativnih rizika. U rad je uključen pregled metodologija i statističkih tehnika koje se mogu upotrijebiti za izračun rizika predviđenih sporazumom Basel, od jednostavnijih prema složenijima, uz razmatranje problematike koja slijedi svaku od metoda. Osim izračuna kreditnog rizika, razmatra se i problematika izrade kredit skoring liste ili ocjena klijentove kreditne sposobnosti koja je usko povezana s problematikom kreditnog rizika i izračunom karakterističnih varijabli vezanih uz procjenu kreditnog rizika (PD, LGD, EAD, M). U rad je uključen i pregled određenih metoda kojima se može procijeniti ispravnost modela te ocijeniti mjere separacije modela. Rad sadržava prijedlog modularne implementacije sustava koji će podržati bankovno poslovanje prema smjernicama Basela II, izračun karakterističnih vrijednosti rizika u raznim portfeljima banke i prijedlog za korištenje tehnologije WebServisa za komunikaciju među modulima. Sustav prati segmentaciju prema strukturi sporazuma Basel i prepoznaje potrebu za implementacijom više statističkih metoda radi dodatne kontrole i validacije internih modela. Predloženi poslovni procesi uključuju raščlambu sporazuma Basel na tri stupa te izvještajnu podršku višem upravnom sloju banke, regulatornog tijela i tržišnih sudionika.Under the guideline that is prescribed by Basel capital agreement, banks will have incentive to build reporting and management system, which will help them in assessment of credit, market and operational risks, by using internal models. This thesis contains an overview of methodology and statistical techniques that may be used for risk assessment and calculation according to Basel agreement, from basic methods to advanced ones. Besides calculation of credit risk, an overview of credit scoring techniques which are closely associated with assessment of credit risk and calculation of characteristic variables important for credit risk assessment (PD, LGD, EAD and M) has been made. Thesis contains an overview of methods that can be used for assessment of model validity and assessment of discriminatory power of the model. General guidelines were made for building a management system under the Basel capital accord guiding principle, a system that can be used in assessment of characteristic risk values through different portfolios. Web Services technology is introduced as communication interface between modules. System is modularly segmented, in alignment with Basel agreement, and the need for implementation of more then one statistical method was recognized. More statistical methods will help in better control and better validation of internal models. Suggested business processes incorporate partition on the three Basel pillars and reporting support for bank management, supervisory authorities and public access