Comparative study of modern credit risk assessment methods
The British University in Dubai (BUiD)
The ability to predict bankruptcy is of great value for investors, lenders and other stakeholders of the companies. Moreover as it has been shown by the recent crisis originated in sub-prime market the financial distress can have multiplicative macroeconomic effect and bring high cost to economy of countries and the society. Thus there are various models to forecast failure of the firms which have been developed and proposed by academics in the recent years. This dissertation presents the basic framework and structure of four credit risk assessment models, namely (1) Merton's structural model, (2) KMV, (3) Z-score, and (4) Binominal approach. Then, work discusses limitations to practical usage of each model, and it also explains some necessary conditions before implementing these models. Real historical data was used to examine the effectiveness of each model in early bankruptcy forecasting. Financial variables of fourteen companies from four different industries were analyzed with two companies in the sample which eventually went bankrupt. The following industries are under consideration in this study − (1) banking, (2) automobile, (3) electronics, and (4) oil and gas sector. Back testing simulation was run on company's financial data collected for 3-5 years pre sub-prime crisis time horizon. On purpose, sample from each industry contains one company that has really defaulted in subsequent years. As a conclusion, work contains a discussion on results obtained to derive a conclusion on which risk assessment model(s) is (are) best in identifying pre-default companies.
Merton, KMV, Z-score, binominal, Distance-to-Default (DD), Expected Default Frequency (EDF), Implied Default Probability (IDP)