Evaluation of BIST manufacturing firm failures within the frame of panel multinomial response models

Qualitative response models are one of the most widely used models in econometric literature. By means of these models, probability of an event to take place or not can be modeled and also probability can be modeled for more than two situations. Models in which dependent variable is described with two categories is called as binomial response models, while models in which dependent variable is described with more than two categories is called as multinomial response models. Usage of these models, especially in medicine, sociology and other human sciences is highly prevalent since possibility of a qualitative phenomenon to happen or not in qualitative response models can be estimated. In this thesis study, by using panel multinomial response models, it has been studied whether there is a relationship between financial successes of firms that operate in BIST manufacturing sector and financial ratios that belong to firms. Primarily, a qualitative dependent variable with three step has been obtained and then by using firms financial ratios as independent variable, pooled multinomial logit, multinomial logit with fixed effects and multinomial logit with random effects models have been predicted respectively. With respect to explaining firm success, the most suitable model is determined as multinomial logit model with fixed effects and financial ratios that explain success are detected. Financially, obtained results live up to expectations.