Analýza regionální nezaměstnanosti pomocí regresních modelů s prostorovou a časovou korelací proměnných

The relations between macroeconomic variables distributed over time and space could be very complex. Therefore, when performing empirical analysis, it is important to take into account as many assumed features of the examined variables as possible. This masters thesis describe advanced econometric methods incorporating spatial and temporal features to effective evaluation and quantification of properties and relationships within examined variables. The main contribution of this paper is introduction of dynamic panel regression model that account for numerous features that one could expect within modelled relationships and provides estimation procedure to obtain consistent estimation of corresponding parameters. Since this model is new to the available literature, custom R implementation of corresponding estimation procedure is provided as well. This model allows for modeling spatial and temporal dynamics, as well as spatially heterogeneous responses while using endogenous regressors in one framework. All described procedures are then applied to unemployment rates in selected European countries.