The adjustment of national education systems to a knowledge‐based economy: a new approach

Abstract

In his article ‘Globalisation, the Learning Society, and Comparative Education’, Peter Jarvis recommends lifelong learning in the period of globalisation as a topic ripe for scholarly research. In particular, he argues for the examination of the extent of lifelong learning around the world and its relation to different levels of employment. This article contributes to this line of inquiry by analysing how education policies facilitate adjustment to economic change and examining how advanced industrialised countries (AICs) compare in their promotion. Principal component analysis is used to construct indices for education systems that reflect these two objectives, and the results reveal considerable cross‐national variation. The Nordic countries appear well‐positioned to cope with changed skill needs. A closer look at the cases of Denmark and Italy portrays how a national education system can facilitate or hinder adaptation, respectively.

Acknowledgements

I would like to thank Evelyne Huber, Anja Jacobi, John D. Stephens, and the editors for helpful comments. All remaining errors remain my own.

Notes

1. Due to data and space limitations, issues such as curricula, informal learning, or teaching methods are not included in the analysis despite their clear relevance.

2. Specifically, through the spectral decomposition theorem, it is possible to rewrite the covariance matrix, (¹/n)XX^T, as the matrix ULU^T. The principal components are then the row vectors of the matrix UT.

3. The spending variable controlled for purchasing power parity.

4. For dissenting views, Ehrenberg and Brewer 1995 and Hanushek 1996.

5. For dissenting views, see Bennett, Glennerster and Nevison (1992), Robinson (1997), and Psacharopoulos (1987).

6. For the remainder of the analysis, I retain only the first unrotated factor for each dimension. I do this for three reasons. The first reason has to do simply with presentation. Comparison is easier if each dimension has only one component. Beyond presentational ease, however, the first component identified for each dimension captures by the most variance among the variables included. A glance at the factor loadings for the variables in each dimension demonstrates that most variables load quite highly on Factor 1, both in general and in comparison to the loadings on Factor 2. Finally, I retain unrotated factors since, as explained in the methodology section, I am interested in summarising the data rather than explaining the shared variance among the variables.

7. Data is from 2006.

8. Data is in US dollars controlling for purchasing power parity.

9. It is perhaps surprising that Danish students do not receive higher scores on average since it ranked so high on the indices for national education systems produced here. First, although the ordering of the countries changes somewhat in Brown, Hesketh and Williams’ measure, the overall distinction between which countries rank high and low is consistent. Second, children in Denmark are less exposed to testing because they do not receive grades for the first nine years of compulsory schooling. Test‐taking ability may therefore influence the results.

10. The data comes from the Continuing Vocational Training Survey II conducted by Eurostat.