Multivariate Extensions to Discrete Choice Modeling

This thesis on econometric modeling consists of two parts.
First, I consider multivariate extensions to univariate discrete choice models. The focus of the chapters is on model representation and parameter estimation methods. Univariate choice models are extended to allow for associations between different univariate choices. When the number of choices involved is large, parameter estimation in these multivariate extensions leads to numerical problems. Therefore, new feasible parameter estimation methods are introduced.

In the second part, I discuss the effect of forecasts on future values of macro-economic variables. Since economic agents react to forecasts, these might affect the course of the economy.