Probability of default estimation for corporate segment in Serbia and development of internal credit rating for the bank's needs

[N. N. Nikolić]

312 listova

Tehničke nauke / Finansijski menadžment

Technical sciences / Financial management

Predmet istraživanja ove doktorske disertacije je razvoj novog metodološkog okvira, kao i praktična implementacija internog kreditnog rejtinga modela za segment velikih i srednjih preduzeća u Srbiji. Osnovni cilj jeste kvantifikacija kreditnog rizika preduzeća u Srbiji, koja je omogućena korišćenjem novo-razvijenog modela kreditnog skoringa, kao i procena boniteta preduzeća, na osnovu uspostavljenog internog kreditnog rejting modela. Podaci na kojima se bazira istraživanje, predstavljaju sedmogodišnje podatke iz finansijskih izveštaja, odnosno bilansa stanja i bilansa uspeha preduzeća u Srbiji za periode posmatranja od 2006. do 2012. godine. Korišćenjem ovih podataka omogućava se kvantifikacija verovatnoće difolta preduzeća u Srbiji, uzimajući u obzir specifičnosti ovog tržišta i poslovnog ambijenta. Na bazi uspostavljene originalne metodologije, od strane autora, u ovoj doktorskoj disertaciji pokrivene su tri velike oblasti. Kao prvo, urađena je analiza velikog broja finansijskih racija, kao i pronalaženje najprediktivnijeg finansijskog racija u Republici Srbiji, na osnovu raspoloživih višegodišnjih podataka. Primenjena je i odgovarajuća transformacija finansijskih racija i korišćenje u svrhu razvoja modela kreditnog skoringa. Kao glavni rezultat istraživanja razvijen je i predložen novi model kreditnog skoringa koji kvantifikuje verovatnoću difolta preduzeća. Ovaj novi model pokazuje veću prediktivnu moć od modela koji su razvijani i korišćeni za inostrana tržišta poput Altmanovog Z-skor modela. Kao konačni rezultat, predstavljen je i uspostavljen interni kreditni rejting model, primenom originalnog pristupa kojim se uobličavaju i koriste rezultati prethodno razvijenog modela kreditnog skoringa. Konačni modeli kao i njihovi rezultati su prošli proces validacije predložen od strane relevantne akademske literature i novih Bazelskih standarda. Izračunati su ponderi kreditnog rizika na osnovu principa Bazela II i III i kroz retrospektivu je pokazano je da su svi rezultati proistekli iz istraživanja dokazali polazne hipoteze ove doktorske disertacije.

The main subject of interest of this doctoral thesis is development of new methodological framework and empirical implementation of internal credit rating model for medium and large companies in Serbia. The main goal is credit risk quantification of Serbian companies by employing newly developed credit scoring model, as well as, the assessment of credit quality of companies by utilizing developed internal credit rating model. The data, which has been used as the basis of this research, is given in form of seven years history of financial statements, in form of balance sheet and income statement, of Serbian companies for the periods from 2006., up to 2012. year. The utilization of this data enables to quantify probability of default of Serbian companies by taking into consideration specific properties of this market and business environment. Based on new and the originally established methodology by the author this doctoral dissertation covers three large areas. The first is represented by the analysis of large number of financial ratios, as well as, a search for the most predictive financial ratio in the Republic of Serbia, based on available data history. Corresponding data transformation of financial ratios has been utilized with a purpose of credit scoring model development. As main result, the new credit-scoring model that quantifies the default probability, has been developed and proposed. This new model has a greater predictive power compared to models, such as Altman Z-score model, which have been developed and used on the foreign markets. As the final result, the internal credit rating model has been proposed and established by applying original approach, which shapes and uses the results gathered from previously developed credit scoring model. The final models and its results have been governed through the validation process which has been proposed by relevant academic literature and new Basel Capital Accord standards. The credit risk weights based on Basel II and III principles have been calculated and it has been demonstrated that all results presented in this research have confirmed the starting hypothesis of this doctoral dissertation.

Kreditni skoring; Interni kreditni rejting; Finansijski racio; Kreditni rizik; Algoritam; Migracione matrice; Bazel II i III.

Credit scoring; Internal credit rating; Financial ratio; Credit risk; Algorithm; Migration matrices; Basel II and III.

UDK=005.5:336.71(497.11)(043.3)

Tekst