This course aims to evaluate credit default prediction on Bosnia and Herzegovina's national banking market and on its constitutional bodies (Federation of Bosnia and Herzegovina and Republika Srpska).

One of the key areas of interest for risk management academics for a long time has been the ability to categorise organisations into various predefined groups or finding an acceptable instrument that would replace human judgement in classifying companies into good and poor buckets.

Logical regression (logit) and multiple discriminant analysis (MDA), two common statistical techniques, were used to examine the likelihood and precision of default prediction and compare the predictive power of each technique. The outcomes demonstrate that the developed models have strong predictive power.

Some variables in logit models have a greater impact on the default forecast than others. With very high regression coefficients, or a considerable impact on the model's capacity to predict default, return on assets (ROA) is statistically significant in each of the four periods before to default.

MDA models produce similar outcomes. Additionally, it is discovered that multiple discriminant analysis and logistic regression have different levels of predictability.

Course Duration:-3h 9m