Individual Earnings, International Outsourcing and Technological Change: Evidence from Italy

Abstract

The aim of this paper is to evaluate empirically the relative effects of international outsourcing of materials and services and of ICT capital deepening on wage inequality between blue and white collars in the Italian manufacturing industry during the period 1985–1999. We merge an administrative data set on workers’ wages and individual characteristics with data on imported inputs from Italian input–output tables and other sector-level variables. Our results confirm that both material and service outsourcing widen the skilled/unskilled wage gap while ICT capital deepening positively affects real wages regardless of the worker's status. However, important differences emerge when the overall sample is split between traditional and innovative sectors.

Keywords:

• International outsourcing
• ICT
• wage inequality

JEL CLASSIFICATIONS :

• F16
• J31
• C23
• O3

Acknowledgements

This paper is part of the project The Evolution of Inequalities in the Italian Labour Market, co-founded by the Italian Ministry of University and Research, prot. 2005132317. The authors thank ISFOL for providing individual workers’ data. They also warmly thank one anonymous referee, Riccardo Lucchetti, Matteo Picchio, Francesco Venturini the participants at the the Conference ‘Innovation, Internationalization and Global Labor Markets’, Centro Studi Luca d'Agliano Fondazione Luigi Einaudi (Turin, 2009), the Workshop ‘The evolution of inequalities in the Italian labour market’ held at Catholic University of Milan (Milan, 2008), and the 21st AIEL Meeting (Udine, 2006).

Notes

¹Machin and Van Reenen (1998) confirm the SBTC hypothesis studying a panel of 7 countries (Denmark, France, Germany, Japan, Sweden, UK, US) over various time intervals (within the period 1973–1989) with 15 manufacturing sectors. More recently, Bratti and Matteucci (2005) point out that the evidence in favor of SBTC for European countries is less straightforward. For further refinements on the relation between technological change and wage inequality see, among others, Aghion et al. (2002) and Borghans and ter Weel (2004).

²They found that the change in outsourcing can account for 30 to 50% of the increase in the non-production workers’ relative wage in the USA manufacturing sectors in the period 1979–1990.

³The result, though, is sensitive to the measure of ICT capital adopted and sometimes bears a smaller effect of the latter variable with respect to the effect of international outsourcing.

⁴Egger and Falkinger (2003) develop a complete characterization of the distributional effects of international outsourcing in the Heckscher-Ohlin framework where the factor and sector bias are reconciled according to a final equilibrium with specialization or diversification. Kohler (2001), instead, proposes a specific factor model allowing for the possibility of a welfare reducing effect of outsourcing for the domestic economy, even without any market distortion.

⁵Another stream of literature addresses the role of outsourcing in enhancing productivity. Amiti and Wei (2009) find a positive effect of outsourcing of materials and services for the US. For Italy, Lo Turco (2007) and Daveri and Jona-Lasinio (2008) confirm the positive role of outsourcing of materials on productivity, while services outsourcing seems to reduce productivity.

⁶Recently, the combination of micro data with sector level offshoring intensities has also been used to investigate the relationship between the latter and workers’ post displacement wages (Crinò, 2010).

⁷Helg and Tajoli (2005) use this very narrow definition of outsourcing, which only conveys information on manufacturing re-imports.

⁸To be more precise, only recently – and subsequent to the publication of the working paper version of this work (Broccolini et al., 2007) – do Geishecker and Gorg (2008b) deal with the wage effects of the outsourcing of services for the UK economy.

⁹Gross wages are the sum of net wages, taxes and social contributions on workers; social contributions on firms are not included in gross wages.

¹⁰ Istituto per lo Sviluppo della Formazione Professionale dei Lavoratori (Institute for Training Workers).

¹¹For a detailed description of the dataset, see Centra and Rustichelli (2005).

¹²This price index is calculated by the Italian Institute of Statistics (ISTAT) with respect to blue and white collar households.

¹³We consider 19 sectors, according to the ateco 81 classification, 2 digits. The dataset is available at http://www.giuri.unige.it/iotables/index.html

¹⁴In the descriptive analysis, the indicators of outsourcing and capital deepening are in levels and not in logarithms, in order to be more readable.

¹⁵The classification of sectors as ‘Traditional’ or ‘Innovative’ follows the one in Lall and Mengistae (2005): the 19 ateco-81 sectors are classified as Innovative according to the existence of economies of scale and to the technological content of their typical activities; the remaining sectors are classified as Traditional.

¹⁶We take the natural logarithm of the share of outsourcing as previously defined and of the other variables introduced below, so that we can easily interpret their estimated coefficient as elasticities. However, we will also check for the robustness of our results using the linear and the logistic transformation of these variables.

¹⁷This results is also consistent with the a priori that, in our specification, the additional hypothesis required by the RE of no correlation between the unobserved effects (i.e. education, innate ability) and the explanatory variables is likely to fail.

¹⁸The results of these tests are available from the authors, on request.

¹⁹The complete estimation includes a set of control variables available in the data set, namely workers’ ages (linear and squared), days worked (linear and squared), regional, yearly, firm size and sectoral dummies.

²⁰Results are not reported for the sake of brevity, but they are available from the authors upon request.

²¹In doing this procedure, we follow Geishecker and Gorg (2008a). Nonetheless, we are fully aware that possible issues of simultaneity and reverse causation would need a deeper inspection and might represent a potential problem for our estimates.

²²As regards industry level controls, sectoral productivity in general is related to an increase in wages, although it is associated with an increase in the wage gap in the innovative sectors. The degree of skill intensity has a limited effect only on wage inequality in traditional sectors, where it raises the daily real wage of skilled workers.

²³Results are not reported for the sake of brevity, but they are available from the authors upon request.