Beyond unobserved heterogeneity in computer wage premiums / Data on computer use in Germany, 1997 – 2001

Beyond unobserved heterogeneity in computer wage premiums: Most findings on the (non-)existence of a wage premium on computer use are biased because they are based on single-equation estimation of a wage equation. Controlling for fixed effects ignores the simultaneity problem. Through the introduction of a latent variable, “PC-feasibility”, we tackle the problem of simultaneity and account for unobserved heterogeneity. Due to the simultaneous nature of wage determination and computer use, the premium for computer use becomes dependent on person and job characteristics. Imposing testable restrictions on the reduced form enables us to identify the factors that determine wages and enhance computer use. The model is estimated using German data, 1997-2001 Data on computer use in Germany, 1997 – 2001: This paper describes the GSOEP data, 1997-2001, used in Muysken and Schim van der Loeff (2006). The data contain relevant information on individuals, with a focus on the computer use both at home and at work. The construction of the relevant data set for the period 1997 - 2001 is presented in section 2. A more detailed discussion of the data is presented in section 3 for the 1997 wave. Interesting observations are that most workers who use a computer at work started to use a computer at home simultaneously or later. Also the average number of years of employment for workers exceeds the period of the sample. Moreover there is a huge amount of inertia for computer use at work. This emphasises that fixed effects do not only control for unobserved individual characteristics, but also firm and job related characteristics. Finally, for rather homogenous groups of workers wages do not appear to vary systematically with computer use.