Discussion Papers, Interdisciplinary Research Project 373: Quantification and Simulation of Economic Processes 2002,13
Using 1985-1999 data from the German Socio-Economic Panel Study (GSOEP) to analyze wages we confirm the hypothesis that existing computer wage premiums are determined by individual ability or other unobserved individual characteristics rather than by productivity effects. While a rather large personal computer (PC) wage premium was found in the crosssectional regressions even after the inclusion of standard controls, the conventional longitudinal regression analysis revealed substantially lower or statistically insignificant coefficients, as have other studies. In addition, a new method of testing the two competing explanations for computer wage differentials against each other was found: future PC variables were employed in the wage regressions in order to obtain a further control for worker heterogeneity. The finding that future PC variables have a statistically significant effect on current wages leads one to conclude that computer wage differentials can be attributed to worker heterogeneity rather than to computer-induced productivity.