ZEF Discussion Papers on Devlopment Policy 7
33
reconcile these differences, the authors construct a computer-investment variable (similar to
those used by other authors) and examine the effect of this variable on the share of non-
production workers. Regressions using these variables suggest that the share of non-production
labor is positively and significantly correlated with computer investments.
Evidence relating to changes in the skill composition
of the work force from other
developed countries is reviewed in Kramarz (1998). A study of 402 plants in Britain suggests
that the introduction of computers in plants is associated with an increase in the share of white-
collar workers at the expense of unskilled, manual workers. Similar results are found in a study
using data from French firms.
Despite the cross-country consistency of these studies (based on firm-level data), the
results regarding the influence of computer use on employment obtained from workforce data
does not display the same pattern of consistency. Card
et al.
(1997)
examine the effect of
computer use on the employment rates of various age-education groups. Based on their
knowledge of the institutional environment in these three countries, the authors expect that the
effect of these technologies would have the greatest negative impact on employment in France,
followed by Canada and then the US (alternatively, computer use should be most strongly related
to the employment rates
of skilled workers in France, followed by Canada and the US).
40
However, the results do not seem to follow this pattern. In the case of the US, the results show
that groups that use computers most intensively record an increase in group employment rates.
In the case of Canada and the female sample in France, there is no significant computer use-
employment relationship. The results from the male French sample are similar to those obtained
for the US. This indicates that despite
the wage rigidity in France, the magnitude of the relative
decline in the employment rates of less skilled workers (which may be explained by the diffusion
of computers) is the same as that in the US.
Concluding remarks:
The evidence reviewed above suggests that, at least in the case of
the US, the increase in the relative demand for skilled labor and the subsequent increase in wage
inequality are related to the spread of ICTs. However, results from
other developed countries
that have had a similar diffusion in computer use do not provide such a clear picture. This
suggests that the outcome in each country/region varies according to the different institutional
features governing the labor market (training possibilities, unions, minimum wages and
collective bargaining). While it is possible that ICTs may enhance the demand for skilled labor,
impeding their diffusion on the basis of distributional concerns seems unwarranted.
40
This hypothesis is based on the idea that, if a similar negative demand shock affects
less skilled workers in all
three countries, then given the labor market flexibility in the United States, the shock should result primarily in a
decline in the relative wages of less skilled workers. In France, where labor markets are relatively inflexible, the
shock should result largely in a decline in the relative employment of less skilled workers.