Credit risk modeling for multilateral lenders

Shirzadi, Nazanin (2015) Credit risk modeling for multilateral lenders. PhD thesis, University of Glasgow.

Abstract

Financial crashes, bubbles, panic in the banking industry, currency crises and even sovereign defaults continue to occur periodically. Therefore, when international or multilateral lenders contemplate on lending credit to customers who are located in different countries, they require a meticulous method of analyzing every aspect to select the best customers, amongst numerous credit proposals from different countries. Moreover, while lending to selected customers, multilateral lenders need to take into account and consider the risk premium in their pricing methodology. Even after having selected sound customers, one should not neglect adequate loan loss provisions in order to safeguard themselves against unexpected changes in financial situations of customers. This may result in credit default. Although several credit scoring methodologies exist for calculating the risk of individuals and corporate customers, most of these methodologies are based on default history and there appears to be a lack of an appropriate methodology when faced with minimal credit default history. Usually, financial institutions and very large corporations are characterized by nil or a very low default history. Following this introduction, this dissertation aims to contribute towards these aspects in the form of three self-contained essays. The first chapter is concerned with determining the main factors, which affect the financial health of financial institutions. More specifically, this is undertaken by employing the two-way panel model and data from financial institutions in several Asian countries. The study attempts to determine bank specific and macro level factors affecting the financial soundness of these financial institutions. In the second chapter by following a similar approach of analysis, this study attempts to detect the main determinants of financial health for very large corporations. These corporations are another group of customers for multilateral lenders. In this case, data from very large corporations in Eastern European countries, which are characterized by their in-transition economies, are employed. Considering the dissertation's findings that are supportive of existing literature, the third chapter addresses the design of two credit scoring/rating models employing fuzzy logic methodology and based upon results from previous chapters. The scoring/rating results of the two models are then analyzed in comparison with the Capital Intelligence rating agency and stock exchange market performance results to assess robustness. This proves the relative robustness of our designed models. Overall, this thesis not only combines and investigates topical issues; moreover, it does so employing various techniques with the intention to contribute on the methodological level. The study is concluded by highlighting policy implications by providing direction for future research.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: Fuzzy Logic, Credit Scoring Models, Privatization, Bank Profitability, Asian Banking Sector, Random Effects, Multilevel Modeling, two-way panel, Central and Eastern Europe, Asian Economies, Fuzzy computation layers, Fuzzy Network.
• Subjects: H Social Sciences > HB Economic Theory
• Colleges/Schools: College of Social Sciences > Adam Smith Business School > Economics
• Supervisor's Name: Cerrato, Professor Mario
• Date of Award: 2015
• Depositing User: Dr. Nazanin Shirzadi
• Unique ID: glathesis:2015-6592
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 06 Aug 2015 14:10
• Last Modified: 13 Aug 2015 07:40
• URI: https://theses.gla.ac.uk/id/eprint/6592