Reindustrialization patterns in the post-socialist EU members: a comparative study between 2000 and 2017.

• Author: Nagy, Benedek

1. Introduction

   Today reindustrialization, and the associated 'fourth industrial revolution', plays a crucial role in economic policies. Reindustrialization itself emerged in the economic literature about three decades ago and following the global economic crisis in 2008 it has again gained attention (Krugman, 1988; Szirmai, 2012; Tregenna, 2013; Cristopherson et al., 2014). Reindustrialization, which we are defining as an increase in the share of manufacturing employment, can both be a spontaneous economic process and a result of an economic policy effort, and it is closely related to the concepts of industrialization and deindustrialization. Manufacturing is a key sector that scholars use to analyze these processes (Szirmai, 2012; Tregenna, 2013). Industrialization primarily refers to a structural change that is apparent in the declining importance of the agricultural sector, an increase in the number of persons employed in industry, and an increase in the share of industry within gross value added (Scott and Storper, 1992; Krugman, 1988; Szirmai, 2012; Tregenna, 2013). By the second half of the 20th century, with the focus on services in developed countries, the significance of industry and industrial policy saw a substantial decline as well as industry's social and political prestige. This process is called deindustrialization (Cairncross, 1982; Kudina and Pitelis, 2014; Pelkmans, 2006; Tregenna, 2009; Wolman et al., 2015). The literature identifies two types of deindustrialization. Mature deindustrialization is characteristic of high-income countries, where manufacturing loses importance with the emergence of the service sector, an increase in the efficiency of industrial activities, and outsourcing lower value-added activities to developing countries (Saeger, 1997, Rowthorn and Ramaswamy, 1997; Alderson, 1999; Rowthorn and Coutts, 2004; Palma, 2005, 2008; Cristopherson et al,. 2014). In developing countries premature deindustrialization (e.g. with the emergence of IT services) is a kind of early collapse of industry (McMillan et al., 2017; Palma, 2005, 2008; Rodrik, 2016; Tregenna, 2016). This deindustrialization is premature because it happens soon after industrialization has started, while the share of industry in the economy is still much lower than in the exemplar countries of industrial revolution (e.g. United Kingdom or Germany). The literature mainly uses the share of manufacturing employment to assess deindustrialization (Tregenna, 2009; 2016) but often both employment share and GVA share is considered (Tregenna, 2013; Rodrik, 2016).

   In order to boost and revitalize the European Union's economy and competitiveness after the crisis of 2008, EU economic policy began to focus on reindustrialization, particularly in post-socialist member states (Lux, 2015; Nagy and Lengyel, 2016; Lengyel et al., 2017; Rachwall, 2015). The EU announced the Europe 2020 Strategy in 2010, aiming at achieving smart, sustainable and inclusive economic growth. To this end, the European Commission--among others--made a proposal on the overall restructuring of the European industrial policy (EC, 2010; EC, 2014a; Landesman, 2015; Torok et al., 2015). Although industrial policy is one of the flagships of the Europe 2020 Strategy (An industrial policy for the globalization era), specific objectives to be attained were not yet formulated at that time. The European Commission's Communication two years later set a clear goal: the share of industry within GDP should increase from the approximately 15 per cent it was then to at least 20 per cent by 2020 (EC, 2012), and simultaneously it should play a major role in economic growth and job creation (EC, 2014b). That is, (re)industrialization has become a crucial element in European economies.

   Analyzing the manufacturing (1) trends in ten post-socialist EU members (PS10) and in the old EU states (EU15), (2) reveals some conflicting trends (Figure 1). In 2017 manufacturing in the PS10 group has a considerably higher share in produced GVA (26.7%, with a standard deviation of 6.8) than in the EU15 group (16.2% with a standard deviation of 6.0). (3) Manufacturing's share within employment is also higher in the PS10 group (20.3%, standard deviation is 3.5) than in the EU15 group (12.3%, standard deviation is 3.0). (4) In addition, the significance of manufacturing exhibits considerable variety between the member states in each group. In a ranking of the share of manufacturing employment in 2017 for all the countries involved in our study, eight of the first ten positions are held by post-socialist countries (with the Czech Republic in first place with 26.7%), while only two of the old member states of the EU rank in the top ten (Germany occupying the ninth place with 17.2%). Among the post-socialist countries the lowest value is in Latvia (13.4%) and even this is higher than the average of the EU15, which is 12.3%. However, the development of manufacturing is not homogeneous even in the case of post-socialist countries (Stojcic and Aralica, 2017), despite the fact that these countries were in an artificially "over-industrialized" condition before the regime change of 1990 (Barta et al., 2008; Chapman, 2013; Lux, 2017). Figure 1 illustrates a fast and substantial decline in the share of manufacturing in the PS10 countries in the late 90s, mostly in terms of employment. This period is still the transition period for the PS10 countries from socialism to capitalism (see e.g. Farkas, 2016), the time of adjustment to the new economic environment that is characterized by opening up to free trade, privatization, high overall unemployment and high inflation. We consider these processes to be different from organic deindustrialization.

   In the EU15, the share of manufacturing within employment decreased since 2000 and its share within GVA stagnated or slightly decreased, with the global financial crisis accelerating the process a little in terms of GVA. These changes fit the more strict requirements of deindustrialization when both shares are required to decrease. In the PS10 states, however, the share of manufacturing in GVA is almost continuously increasing, while the share of manufacturing employment seems to be decreasing. This trend does not meet the above-mentioned strict definition of deindustrialization. In light of the inconclusive trends we investigate in this paper whether deindustrialization or reindustrialization takes place in the PS10 and the EU15 countries and compare the evolution of the manufacturing sector of these country groups. De- and reindustrialization is generally assessed with reference to the change in the share of manufacturing employment and gross value added, and also to the change in their absolute levels. As any combination of direction of changes in these variables is possible for our study we define reindustrialization as an increase in the share of manufacturing employment and deindustrialization as a decrease therein (Tregenna, 2016). In order to analyze and compare the performance and evolution of the manufacturing sector of the two country groups mentioned above we use the number of persons employed and gross value added as basic variables. Applying a decomposition method, we analyze the change in labor intensity (essentially, labor productivity) and growth of manufacturing GVA.

   In the first part of the study we present the methodology we apply to examine reindustrialization, followed by the analysis of the characteristics of manufacturing based on this method between 2000 and 2017 in the EU15 and PS10 countries, in three separate time periods. In the last section of our study we discuss the results regarding reindustrialization and the questions left open.

2. Methodological background

   In this section, we examine the direction of processes taking place in the manufacturing sector of the EU15 and PS10 countries. We examine ten post-socialist countries (PS10), out of which eight (the Czech Republic, Estonia, Poland, Latvia, Lithuania, Hungary, Slovakia and Slovenia) joined the EU in 2004, and the other two (Bulgaria and Romania) did so in 2007. From the PS10 group we are excluding Croatia, first because its EU accession happened much later (2013) and second because as a consequence of this the Eurostat data about Croatia is very scant.

   The performance of the manufacturing sector is generally assessed by various decomposition methods in the field of industrial economics. These methods concentrate on one of the followings as output variable: (1) number of persons employed in the industry; (2) gross value added of manufacturing; (3) labor productivity, and they use more or less detailed sectoral breakdown of the economies analyzed. For example, Tregenna (2009, 2013) uses the decomposition to separate the effects of factors causing a change in employment, Cantore et al. (2017) concentrate on gross value added, and McMillan and Rodrik (2011) are decomposing the change in the labor productivity. All of these methods look at the same phenomenon from different angles. There exist also other methods like the shift-share analysis or indices such as the Lilien index, the Havlik index, or UNIDO's index of structural change but they are generally used to study multiple sectors in either a single country or for a variety of countries. Since we are concentrating on one single economic activity (manufacturing) in different countries, the use of decomposition to answer our research question is more appropriate.

   In our analysis we follow the three-way decomposition method used by Tregenna (2009, 2013) which separates the effects of change in manufacturing labor intensity, share of manufacturing in GVA and overall labor productivity on the change in the share of manufacturing employment. The employment share of manufacturing can increase through a decreasing labor intensity in manufacturing as well as a higher share of GVA being produced in manufacturing or aggregate...