Footnote:
Diese Datenquelle enthält auch Bestandsnachweise, die nicht zu einem Volltext führen.
Description:
The superiority of full information approaches when estimating a system of equation is well known for large samples. However, less is known about the small sample properties of these estimators relative to limited information approachs. This is especially true for the context of Panel data although they are meanwhile a standard tool in estimating static and dynamic economic processes. Here only extensive Monte Carlo simulation evidence is given for single equation models. We expand the scope of Monte Carlo simulation evidence to system estimation. We especially aim to contribute to the methodological debate about appropriate dynamic panel data estimators beside standard - large N, small T - panel data assumptions. Since frequently used dynamic panel data models in 'first differences' have shown a poor empirical performance in these circumstances, we put a special emphasis on appropriate estimators with variables in levels based on Hausman-Taylor (1981) type estimator. We compare the model's performance in terms common evaluation criteria (such as bias and root mean square error) relative to various alternative specifications including the Anderson & Hsiao (1981) IV, the Arellano & Bond (1991) and Blundell & Bond (1998) GMM estimators. Building upon the small sample simulation evidence, the paper then applies a simultaneous equation approach to analyse the effects of regional equalization policy instruments on regional economic output and endogeneized private and public factor demand equations for Germany. By doing so we are able to identify the direct and indirect policy channels of regional equalization schemes, which are often left unidentified in single equation estimations. To give an example, though our empirical analysis strongly confirms recent findings of a negative direct output effect of the German interstate/federal fiscal equalization transfer scheme (Länderfinanzausgleich) via its 'horizontal' (state-to-state) element, the additional analysis of the structural factor input equations ...