Modelling the Confidence in Industry in Romania and other European Member Countries Using the Ordered Logit Model

The application of qualitative choice models is usually made by neglecting the analysis of autocorrelated and heteroscedastic errors. In the current paper, we aim to evaluate and mitigate the effects of violation of such a hypothesis using as example the modeling of confidence in industry in relation to the macroeconomic indicators for six countries of the European Union. The ordered Logit model identified in the paper revealed the common macroeconomic factors which explain the formation of confidence in industry for the countries considered in the analysis. By mitigating the heteroscedasticity problems and specifying in the model the functional form of the error dispersion, the statistically significant improvement of the model performance was obtained.