TRADE INTEGRATION, OUTSOURCING AND EMPLOYMENT IN AUSTRIA: A DECOMPOSITION APPROACH

Abstract

Outsourcing and trade integration of advanced countries is debated with respect to employment effects, in particular for low educated workers – at least in relative terms. We study the employment effects – differentiated by educational attainment levels – of changes in the patterns of trade integration and outsourcing in the Austrian economy over the periods 1995–2000 and 2000–2005 using hierarchical decomposition analysis based on deflated input–output tables. Outsourcing is modeled as changes in the shares of domestically produced intermediates in total intermediates. A similar decomposition of the final demand vector allows us to draw conclusions on the overall employment effects of trade integration. The results suggest that the expected negative employment effects of outsourcing and rising import penetration have been overcompensated by increasing exports. Thus, the overall employment effects of Austrian trade integration have been positive for all educational attainment groups. However, whereas the total effects have been strongest for medium and high educated workers over the period 1995–2000, employment of low educated workers have been strongest and positively affected over the period 2000–2005. This pattern can be explained by a more sluggish export performance together with stronger negative effects of outsourcing and import penetration in medium and high-skill intensive products.

Keywords:

• Outsourcing
• Employment
• Hierarchical decomposition
• Input–output modeling

Acknowledgments

An earlier version of this paper was written in the framework of the project ‘FIW – Research Centre International Economics’ financed by the Austrian Federal Ministry of Economics and Labour. Robert Stehrer also acknowledges funding of this research within the Seventh EU-framework project ‘WIOD: World Input–Output Database: Construction and Applications’ (www.wiod.org) under Theme 8: Socio-Economic Sciences and Humanities, Grant agreement no. 225 281, which enabled the finalization of this much extended version. The authors thank Professor M. Landesmann (wiiw), Prof. M. Luptáčik (IWI), and Josef Richter for useful comments and suggestions. Two anonymous referees provided useful comments to a previous version of the paper.

Notes

¹ Some recent studies focus on growth or productivity effects. Oosterhaven and Hoen (1998) apply decomposition analysis to disentangle the effects of various demand components including exports on real value added growth using inter-country input–output tables for six European countries. Dietzenbacher et al. (2000) decompose labor productivity growth using the same database. ten Raa and Wolff (2001) study the effects of service outsourcing and productivity in the US using input–output analysis. For an overview of current uses of the input–output framework in an international context see Wixted et al. (2006); OECD (2007).

² As an example, in a growing economy with no structural changes we would find positive employment effects of exports but zero effects of outsourcing.

³ This is based on a definition of ‘polar decomposition’ as a relative term, i.e. two decompositions are polar to each other if the corresponding permutations are inverse to each other. Dietzenbacher and Los (1998) use an absolute concept, stating that there are two polar decompositions that are given by the original ordering and its reverse. However, this concept is treacherous since there is actually no ‘original ordering’. Any model can be reformulated as to induce the desired ordering. For example, by employing the transpose operator, becomes . See also De Haan (2001), who addressed this issue.

⁴ This point becomes obvious when we consider that (using only one of two alternative decomposition formulas) .

⁵ A second typology used is the distinction between energy, material and services inputs.

⁶ See UNESCO (2006). This classification comprises seven stages of educational attainment: Pre-primary education (ISCED 0); Primary education or first stage of basic education (ISCED 1); Lower secondary or second stage of basic education (ISCED 2); (Upper) secondary education (ISCED 3); Post-secondary non-tertiary education (ISCED 4); First stage of tertiary education (ISCED 5); Second stage of tertiary education (ISCED 6).

⁷ A referee pointed out that there is a correspondence between educational attainment categories (ISCED) and occupational clasification (ISCO). For example, OECD (2007) suggests the following correspondence (from ISCO 1-digit) to the ISCED categories mentioned above: ISCO 1 (Legislators, senior officials and managers), 2 (Professionals), and 3 (Technicians and associate professionals) corresponding to ISCED 5 and 6; ISCO 4 (Clerks medium Skilled), 5 (Service workers and shop and market sales workers), 6 (Skilled agricultural and fishery workers), 7 (Craft and related trades workers), and 8 (Plant and machine operators and assemblers) to ISCED 3 and 4; ISCO 9 (Elementary occupations) to ISCED 0,1,2; ISOC 0 (Armed forces) has no assignment. Thus one might interpret the results presented below also in terms of occupational categories. However, one should note that this correspondence is less clear when looking at actual data (e.g. Labour Force Survey statistics). Thus, we postpone examining the effects of internationalization by occupational categories to future research.

⁸ A short description of these deflation procedures is provided in the appendix.

⁹ A more detailed analysis would require data to split exports and imports with respect to destination and origin countries which is aimed at an ongoing EU Framework Project (see www.wiod.org). Further, it would imply that the decomposition tree would become even more complex. This will thus be an interesting avenue for further research.

¹⁰ Detailed results including minima, maxima, ranges and standard deviations are available upon request.

¹¹ The first typology follows Peneder (2007), although this refers to industries rather than products. The second distinction of products follows the classification as used in the EU KLEMS database (see www.euklems.net). The classifications are listed in Appendix Table A1.

¹² Even more results are available upon request.

¹³ Note that these figures are thus not directly comparable to employment changes by persons or hours worked.

¹⁴ This severe change in labor input coefficients in agriculture is thus mainly caused by a methodological revision (concerning self-employment) rather than reflecting actual productivity growth and thus should not be given too much attention. One should further note that this has only negligible effects on the other results of the decomposition analysis.

¹⁵ These results are quite robust with respect to different inflation procedures; results are available upon request.

¹⁶ This should, however, not be confused with exports of intermediates, i.e. outsourcing from abroad, which we cannot capture directly – however they are included in exports.

¹⁷ Given the low number of energy products, we present the outsourcing and insourcing effects only for material and service products.