EN
In analyses based on time-series - cross-section data (TSCS) the following problems are very often encountered: groupwise heteroscedasticity, cross-sectional correlation and autocorrelation of the error term, the latter being common for all units or unit-specific. There are different ways of taking those phenomena into account while hypotheses testing and estimating a model. The estimation methods, that are most often used are FGLS (adapted for this case by Parks) and OLS with panel corrected standard errors, PCSE proposed by Beck and Katz. Both these methods are applied to the analysis of labour demand by (chosen) PKD sections and their effectiveness is examined.