|
ZEF Discussion Papers on Devlopment Policy 7
Pdf ko'rish
|
| bet | 30/43 | | Sana | 20.05.2024 | | Hajmi | 363,42 Kb. | | #245457 |
Bog'liq zef dp07ZEF Discussion Papers on Devlopment Policy 7
29
In his work, Krueger appears to conclude that there is a causal link between computer
usage and the wage premium. Whether this is true or whether the computer-use wage premium
is a result of unobserved worker heterogeneity, i.e., workers with attributes that lead to higher
wages are more likely to use computers on the job, is currently a subject of dispute. To examine
this issue in greater detail, Di Nardo and Pischke (1997) use data from West Germany (three
cross-sections of the Qualification and Career Surveys conducted in 1979, 1985-86 and in 1991-
92) to examine the same question as Krueger.
Despite the institutional differences between the US and the German labor markets,
patterns of computer use and the computer-use wage premium are similar in the two countries.
The computer-use wage premium shows a steady increase from 11.8 percent in 1979 to 16.9
percent in 1985-86 and 18.6 percent in 1991-92. In an attempt to shed light on the causal link
between computers and wages, the authors now examine the wage returns on other office tools.
Surprisingly, they report that the returns on pencil use are of the same magnitude as that for
computers. It is difficult to believe that the use of pencils renders individuals more productive.
A more appropriate interpretation is that the positive returns on pencil use reflect selection
effects. The authors argue that if this type of selection is important for pencils, it may be equally
important for the computer effect. Accordingly, the authors conclude that the computer-use
wage premium is largely the result of unobserved worker heterogeneity.
Using cross-sectional data for the years 1987, 1989, 1991 and 1993 and a similar
framework, Asplund (1997) examines the impact of computer use on wages in Finland. The
results reveal that over the years, there is a steady erosion of the computer-use wage premium.
The premium declines from 8.76% in 1987 to a negligible effect in 1993. Dividing the
computer-use variable into six categories based on the type of computer used also reveals a
similar pattern. One reason for this particular pattern may lie in the spread of computer use.
Compared to other developed countries, Finland has more widespread computer use. If the
selection story is true, then as computer use spreads and less skilled workers gain access to
computers, the computer-use wage premium should decline. While this story may be true in the
Finnish case, it is at odds with the results in Germany and the US, where the wage premiums
have increased in spite of the diffusion of computers.
Further evidence based on more comprehensive cross-sectional and panel data from
France (French Labor Force Surveys conducted in 1985 and 1987) is provided by Entorf and
Kramarz (1997, 1998). Their data allows them to construct three computer indicator variables
and three “experience with computer-use variables”. The three computer-use categories are
constructed on the basis of the type of ICT involved (i.e., whether the technology is an office or a
manufacturing technology) and the amount of autonomy left to the user. These categories
correspond roughly to the educational/skill levels of the workers. The first two categories are
computer-based NT (new technology) involving high and average degrees of autonomy,
respectively, and the third is production-based NT involving a low degree of autonomy. Based
on a regression similar to those described above, the authors find that significant wage premiums
|
| |
