Polarizacija na trgu dela: Je kriva tehnologija ali globalizacija?

Če slučajno koga zanima, pri OECD je bil objavljen nov delovni zvezek z našo raziskavo “Labour Market Polarization in Advanced Countries: Impact of Global Value Chains, Technology, Import Competition from China and Labour Market Institutions” na aktualno temo izginjanja srednjega razreda, ki bo naslednje leto (v nekoliko dopolnjeni verziji) objavljen v reviji Industrial and Corporate Change.

Koen Breemersch, Jože Damijan & Joep Konings

The past two decades have witnessed new global trends that shaped the economic fabric of developed economies and transformed their labor markets. One such trend is rapid technological change and the threat it poses to routine jobs through mechanization of the production process and computerization. A second trend is globalization which has manifested itself in the rise of new foreign economic powers, in particular China, and has led to increased import competition in developed economies and the restructuring of production processes in global value chains (GVCs). The emergence of China and GVCs in international trade and the swift changes in technology have led to a reexamination of the traditional way of thinking about the impacts of trade and technological change on labor markets. For instance, the ILO (2015) reports that one in five workers are estimated to work in global value chains indicating increased fragmentation of production into different activities and tasks, which is likely going to affect the relative demand and wages for skilled and unskilled workers. Similarly, Brynjolfsson and McAfee (2014) predict that in the second machine age the growth in productivity has been decoupled from jobs and income as in the digital economy a set of goods and services can be provided at a cost that is often close to zero. New technology does not inevitably reduce the overall demand for labor, but shifts demand to different kinds of work. In this context, labor markets of advanced countries have experienced substantial job polarization, with employment ‘polarizing’ into relatively high-skill, high-wage jobs and low-skill, low-wage jobs (Autor et al., 2006; Goos et al., 2009).

This U-shaped polarization pattern of the labor market is a widespread phenomenon in advanced countries. Although there is a common pattern of employment growth visible across the United States and the European Union, there are substantial differences between countries. This suggests that there is no single factor or common cause at work, but several factors that affect the shape of the labor market development, as also pointed out by Autor (2010). While job polarization was first attributed to skill-biased technological change (Autor and Katz, 1999), recent research has shifted the focus towards offshoring of routine tasks through GVCs (Oldenski, 2014), Chinese import competition (Autor et al., 2013; Keller and Utar, 2016) and to labor market institutions. Acemoglu and Autor (2011) provide evidence on how skill-biased technological change can influence the productivity of various skill groups in performing tasks in production and, hence, affect relative wages and employment shares of these skill groups. Goos et al. (2014) focus on ‘routine biased technological change’ and ‘offshorability’ of such routine tasks. The argument is that routine tasks, like those operating assembly lines, are being progressively offshored to low-wage countries, which diminishes demand for labor in the middle of the wage distribution. The model of Acemoglu and Autor (2011) can also account for ‘offshorability’ of tasks, which has similar effects as skill-biased technological change[1].

So far, most work has focused on the technological channel triggered by innovation and automation (e.g. Autor, 2015). Goos et al. (2014) show that the main effect comes from ‘routine biased technological change’, while offshorability of routine tasks has also contributed to polarization. However, Goos et al. use a subjective indicator taken from surveys to capture the ‘offshorability of tasks’ rather than a trade-based measure of offshoring. Moreover, their measure does not vary over time, which arguably might be important when analyzing polarization. Oldenski (2014) analyzes the impact of offshoring on polarization in the United States using a newly constructed measure, based on the total sales by a foreign affiliate of a U.S. multinational as a share of its total sales. While not much of an effect is found on average wages and employment, significant effects of both offshoring and technological change (proxied by the use of ICT) are found on polarization patterns in the U.S.

The role of Chinese imports in the decline of aggregate manufacturing employment in developed countries and the shift towards services has been receiving increased attention in recent years. Since the entry of China into the WTO in 2001 until the crisis in 2008, U.S. and EU manufacturing employment declined by 3.5 and 3.6 million units, respectively. Autor et al. (2013) estimate that about a quarter of the aggregate decline in U.S. manufacturing employment is due to the rise of Chinese import penetration. The findings of Donoso et al. (2014), Dauth et al. (2014) and Balsvik et al. (2013) confirm that the Spanish, German and Norwegian local labor markets that are specialized in industries competing with Chinese imports, underwent a similar fate. While it is clear that globalization has had an important impact on this process of deindustrialization, not much evidence exists on how globalization, deindustrialization and polarization are related. In a recent contribution, Keller and Utar (2016) linked the rise in Chinese imports to the decline in Danish middle-paid manufacturing jobs which has further aggravated employment polarization. As workers are pushed out of middle-paid employment in manufacturing, they transfer to low-wage services or high-wage employment. These effects hold when controlling for offshoring and technology, suggesting that Chinese import competition operates separately from these channels and is an alternative force of labor market polarization. Overall, the estimates suggest that Chinese import competition accounts for about a fifth of total mid-paid employment decline in the Danish national labor market.

In this paper we use a large sample of European countries to analyze the relative importance of globalization and technological change in explaining this pattern of polarization that has swept developed economies. We also explore whether differences in labor market institutions can help explain the discrepancy in polarization patterns between countries. We make a number of contributions to the growing literature on labor market polarization, globalization and technological change.

First, we rely on a new measure to capture the effect of offshoring of routine tasks. We use trade in value added (TiVA), which is a measure of integration in GVCs, to analyze its impact on job polarization. The TiVA database relies on inter-country Input-Output tables to construct measures of the value added content of trade (OECD, 2015). These new trade measures have already helped in uncovering a number of important new facts: value added exports range from 50 to 90 percent of the value of exports, manufacturing trade is relatively smaller when measured in value added terms and the gap between gross exports and value added exports is quite heterogeneous across countries (Johnson, 2014).

Second, we analyze job polarization in a broad range of 19 European OECD countries, including a number of Central-European countries. While a focus on wage polarization would allow one to analyze how the different forces have affected inequality, we exclusively focus on how the forces are linked to polarization of employment in high- and low-skilled jobs.

Third, we simultaneously explore the effects of offshoring, technology and Chinese import competition in a large cross-country sample rather than at the individual country level.

Fourth, we explore heterogeneity in labor market polarization between countries. In particular, we analyze how labor market institutions may affect polarization patterns. These institutions could affect the relative wages of different skill groups. Therefore, we would expect that the impact of both technological progress and the emergence of GVCs may be different depending on the type of regulations and wage setting institutions that prevail in various countries. To perform this analysis we analyze both manufacturing industries and private industries other than manufacturing. This allows us to tease out various relationships between the degree of involvement in GVCs, technological change, Chinese import competition, institutions and polarization at the industry level.

Our main findings can be summarized as follows: First, polarization is a phenomenon that is predominantly driven by polarization within individual industries. The reallocation of employment away from lowly polarized industries with relatively more mid-skill jobs, such as manufacturing, towards highly polarized industries with relatively more low- and high-skill jobs also contributes.

Secondly, of the three forces we consider to explain polarization within individual manufacturing industries, technological change is the most important one, while we also find some tentative evidence that Chinese import competition also contributed. For industries outside of manufacturing, offshoring and ICT induced technological change are the most relevant.

Thirdly, we find evidence that corroborates the findings in the literature that Chinese net import competition is associated with the decline in employment in less polarized manufacturing industries. For industries outside of manufacturing there are some indications that those industries that have adopted ICT have experienced employment growth.

Fourthly, labor market institutions generally perform weak in explaining the heterogeneity in employment polarization patterns between countries. Since we do not have data on wage polarization, our findings do not exclude that labor market institutions are relevant in affecting wage polarization. Nevertheless, our results suggest that there is ample scope for policies such as skill-development programs, activation policies and sufficient social protection to facilitate smooth transition processes for affected workers.

[1] Another relevant factor in potentially driving polarization is the existence of wage rigidity. This could have made certain skill groups more vulnerable to the effects of technological change and globalization and thus acted as an aggravating factor.

%d bloggers like this: