Employment for Sustainable Development: Sectoral Efficiencies in EU Countries

Mainstream economics perceives labour productivity as the main problem in understanding economic evolution. It is considered as a driving force of economic growth and is often analysed in terms of convergence or national competitiveness (e.g. Demeter et al. 2011; Misiak et al. 2011; Piętak 2014). Generally, productivity is defined as the relationship between outputs and inputs, and in economics, it is considered as the relationship between production and the factors of production. Against this background, labour productivity is equal to the ratio between a volume measure of output (gross domestic product (GDP) or gross value added) and a measure of input use (the total number of hours worked or total employment) (Freeman 2008). As many economists stress, labour productivity depends on the availability and quality of labour resources, applied technologies (Auzina-Emsina 2014), capital accumulation (Ceccobelli et al. 2012) and innovation (Kurt and Kurt, 2015). We assume that the sectoral structure of an economy constitutes another important factor influencing labour productivity. Some sectors are able to generate better results than others and thus labour shifts can either increase or decrease the general efficiency of a country.

Nevertheless, labour productivity is usually analysed at an aggregate level of the economy and the number of studies concerning intersectoral differences in labour productivity is limited. In this relatively neglected field of research the recent studies by Bernard and Jones (1996), Maudos et al. (2000), Carree et al. (2000), Lee (2005), Wong (2006), Badunenko and Romero-Ávila (2014, 2015), O’Leary and Webber (2015), Martino (2015) or Kosmalski (2016) are an exception. They have their roots in classical research on structural development started in the 1930 s by Fisher (1935), Clark (1940) and next by Fourastié (1949) in their three-sectoral concepts of socio-economic development. The search for a sectoral growth pole was present in the works of Perroux (1955), in Hirschman’s (1958) ideas of an unbalanced development, as well as in Rostow’s (1971) studies of economic development. This disaggregated attitude led researchers to compare the economic performance of different parts of the economy. The sectoral approach is adopted in this study, where we consider the efficiency of four sectors: agriculture, industry, market services and non-market services.

Another important issue connected with analyses of labour productivity is specified by the definition of outputs, which can be defined as the desired results of a labour engagement expected to increase national welfare. The mainstream of research that covers international comparisons of levels of development is focused on its economic dimension specified by GDP per capita. Although production and income, however, are undoubtedly the most important determinants of the general level of development, there are other essential factors strongly influencing the welfare of nations. In a framework of sustainable development (SD), these are specified as social and environmental dimensions and complement the economic dimension (e.g. Galdeano-Gómez et al. 2016; Islam et al. 2003). A broad concept of an integrated holistic development assumes that different dimensions of human nature must be balanced and man cannot be viewed solely as homo oeconomicus (Woźniak et al. 2015).

In the literature, as well as in a political debate, SD is concerned with finding a way whereby human socio-economic needs can be met in a harmony with environmental issues, given their strong interdependence (Bolis et al. 2017). The economy view in SD is geared mainly towards improving human welfare, primarily through increases in goods and services consumption. The environmental domain focuses on an integrity protection and a resilience of ecological systems. The social domain emphasises an enrichment of human relationships and an achievement of individual and group aspirations (Islam et al. 2003). The concept of SD covers the issue of intergenerational balance in human well-being described as ‘meeting the needs of the present without compromising the ability of future generations to meet their own needs’ (Our Common Future 1987). It adopts an inclusive wealth theory that assumes that development is sustainable if society’s productive base per capita does not decline over time, and the ability of future generations to meet their own needs is not disrupted (Collins et al. 2017).

In this vein, it is stressed that it is complexity what is required when measuring well-being (Maggino and Zumbo 2012). Nevertheless, because of some essential problems with an operationalisation of SD goals, there still exists a research gap in an assessment of labour productivity described by results of human resources engagement for different spheres of development. Thus, discussing issues of SD we adopt a three-dimensional approach and simultaneously consider social, economic and environmental objectives shaping employment structures. Our study connects a labour productivity issue with the concept of SD and perceives productivity as an indicator of the development.

The aim of the paper is to identify the most efficient structures of employment that are able to generate positive results in the three dimensions of SD. The comparisons are made considering European Union (EU) countries as relatively homogenous economies in terms of socio-economic development and its institutional conditions. In order to indicate the most efficient countries within the EU that can be perceived as structural benchmarks, as well as to specify intersectoral differences in the social, economic and environmental effects of employment, a non-radial data envelopment analysis (non-radial DEA) method was used.

The DEA method allows a relative efficiency assessment of selected decision-making units (DMU) on a background of reference units in a given area of an analysis. In this paper, it is the productivity of employment of selected EU countries. The advantage of the DEA usage stems from its feature as a method that does not require the specification of any particular functional form of relationships between multiply outputs and inputs as well as a money value to aggregate an information. Moreover, this method advantage is connected with the possibility to avoid a problem of the aggregate level of analysis. As different sectors may characterise different levels of efficiency in a process of SD, an aggregate approach may significantly influence the results. The use of non-radial DEA allows to estimate not only the overall productivity of employment, as an average of the productivity of each economic sector i.e. agriculture, industry, services (market and non-market), but also to define necessary shifts of labor resources within sectors. Because the use of appropriate instruments of economic policy may have a great impact on the allocation of labor resources within sectors and shifts of resources from less efficient sectors to those with higher efficiency, therefore the study uses the input-oriented DEA model. Its application allows the assessment of both the general three-dimensional efficiency of employment structures among EU countries and the separate analysis of each dimension of SD, concerning all employment inputs. In this paper, we test the following research hypothesis:

The rest of the paper is organised as follows. Section 2 presents a literature review on interrelations among dimensions of SD in the context of overall productivity improvement. Section 3 briefly describes the research method used, i.e. radial and non-radial CCR model and a concept of canonical correlations as a basis for variables selection. Section 4 presents the first part of research findings, namely data and variables, descriptive statistics, canonical correlations results, as well as a general condition of labour markets in the EU countries. Section 5 presents and discusses an overall efficiency of employment in the examined sectors and the detailed efficiencies namely social, economic and environmental of the studied EU’s member states. The concluding section ends the paper with summaries, general findings, and practical recommendation.

Undertaking research in the field of sectoral employment efficiencies, taking into account economic, social and environmental areas of SD, required a double-track approach to the literature review. Firstly, in terms of a role and importance of these areas for the growth and economic development. Secondly, in terms of a role and significance of a productivity, as one of the features of SD strengthening convergence of regions and countries. Among the referenced works, there are both review papers and papers attempting to empirically assess productivity, in particular, the productivity of employment, including works using DEA method.

Strong interrelations among the SD dimensions are highlighted by many researchers. A starting point for such research can be found in the works of Kuznets (1955), who formulated a hypothesis about a relationship between economic development and income inequalities in a form of an inverted U-shaped curve (the Kuznets curve (KC) phenomenon). The KC is traditionally explained by Lewis’s two-sector model, in which a shift from agriculture to industry decreases inequality. In a phase of digitalisation, however, some authors point out that the disparities grow again (Kiatrungwilaikun and Suriya 2015). Moreover, research in the developing world has proved that the relationship between inequalities and structural change may differ depending on the different phases of structural development (Andersson and Palacio 2016).

Grossman and Krueger (1995) established that the same inverted U-shaped pattern describes a relationship between economic development and environmental degradation (the environmental Kuznets curve (EKC) phenomenon). Concerning the EKC, structural changes specified by industrialisation are associated with environmental deterioration, while increasing de-industrialisation is characterised by lower environmental degradation. Structural changes from agriculture-based economies to industrial and services ones are specified as a basic explanation for the EKC hypothesis by Charfeddine and Mrabet (2017).

In fact, the KC and EKC relationships are not fully validated and some authors neglect their existence in the specified form (e.g. Dinda 2004; Kiatrungwilaikun and Suriya 2015; Nasr et al. 2018). Explanations for both, however, are connected with sectoral changes within the economy. They are stimulated by differences in the sectoral efficiency of resource usage, as well as changing social preferences in consumption and a new lifestyle. Thus, one of important issues specifying development lays in the productivity understood as a ratio between outputs and inputs.

As the main issue connected with the development is an increase in productivity, many authors analyse an aggregate production efficiency and its changes over time. They try to find explanations for a convergence between national and regional economies. Ahmed and Krishnasamy (2013) conducted DEA for productivity changes in ASEAN countries to found its main sources. Ko Li and Zhao (2015) explored a DEA model to discuss the possible strategies of Chinese provinces to improve their competitiveness, which the authors strongly connect with real GDP per capita. Gitto and Mancuso (2015) used a DEA model to decompose changes in labour productivity in Italy in the context of regional differences.

The analysis of labour productivity growth, however, is typically conducted by adopting an aggregate approach to different kinds of economic activity. Some exceptions include work conducted by Maudos et al. (2000), who, while assessing the efficiency of production and regional convergence in Spain, used a 5-sectoral disaggregation specifying: agriculture, industry (excluding energy), energy, construction, and services. As a result, they differentiated between two components of efficiency in aggregate production: intra-sector efficiency and efficiency of the composition of production. Badunenko and Romero-Ávila (2015) proposed a similar approach to sectoral disaggregation. They attempted to discover the sources of labour productivity growth in 5 sectors in Spain i.e. agriculture, energy, manufacturing, construction, and services. Both Maudos et al. (2000) and Badunenko and Romero-Ávila (2014, 2015) came to a conclusion that there are essential differences in sectoral performance. While Badunenko and Romero-Ávila (2014, 2015), however, claimed that aggregate productivity growth is mainly driven by intrasectoral productivity dynamics rather than structural change (sectoral shifts), Maudos et al. (2000) underlined greater gains in composed efficiency as a consequence of structural changes. O’Leary and Webber (2015) presented research results for European regions that rehabilitate the importance of structural change for growth and convergence. They asserted that inter-sectoral structural change appeared to be especially important for those regions situated towards the higher deciles of distribution.

Simultaneously, researchers are trying to specify which sectors are the main drivers of modern development and convergence between nations and regions. Wong (2006) showed that while productivity growth in services and agriculture contributed significantly to OECD convergence, the contributions from employment shift and productivity growth in manufacturing were not statistically significant. Lee (2005) pointed out that for the productivity growth of the Korean economy there is a great need for productivity improvement in services. The above results are in line with Bernard and Jones (1996), who examined the role of sectors in aggregate convergence for 14 OECD countries. Their major finding showed a driving force of services and the diminishing effect of manufacturing in the aggregate convergence process. In the same vein, Carree et al. (2000), while analysing labour productivity across manufacturing industries in 18 OECD countries, showed large inter-industry differences in the extent of convergence. These were explained by a differentiation in a level of average labour productivity. The labour productivity dynamics of the EU during 1991–2007 were investigated by Martino (2015). Despite the generally conditional character of the convergence process, he revealed an unconditional convergence process for financial and business-related market services. Such a convergence was not confirmed for manufacturing and aggregated productivity. A regional analysis of convergence in Poland was presented by Kosmalski (2016). Using DEA approach he compared labour productivity growth among agriculture, industry and service sectors, indicating that their sources were mainly rooted in a technological progress and a capital accumulation. The author stressed that a structural modernisation of the converging economy is important, as it allows the reallocation of labour into the most efficient sectors. The importance of labour productivity was also stressed by Tarancón et al. (2018) who presented the results of a panel data analysis from 1995 to 2011. They examined a relationship between labour productivity and a degree of production efficiency for 24 EU countries. Their results based on the DEA and a regression analysis confirmed the existence of a direct and significant relationship between labour productivity and production efficiency. According to them the observed relationship is also relevant from a policymaking point of view. The context of structural relations stresses a relevance of the service sector as the one contributing to the improvement of an overall efficiency in economy.

Existing literature provides a broad number of results on an efficiency assessment of the individual industry, service or agriculture sectors based on the DEA method. For example, Nazarko and Chodakowska (2015) used the method to examine the construction sector productivity in 25 European countries. Their analysis covered 2006–2012 period and revealed huge differences in productivity of construction industry within the EU countries. DEA method was also used to estimate the efficiency of a power generation sector in the USA (Halkos and Polemis 2018). Apart from the efficiency assessment, the study examined a relationship between environmental efficiencies and economic growth in attaining sustainability. Djordjević et al. (2018) employed a non-radial DEA model for an evaluation of railway efficiency of European countries regarding a level of safety at railway level crossings. Laurinavičius and Rimkuvienė (2017) using DEA provided results on the analysis of agricultural sectors efficiency of 28 EU member states (for 2010–2014). According to their results the agriculture sector of the “new” EU countries is still poorly effective, even though the efficiency tends to increase.

Despite numerous research results on the efficiency of individual branches, sectors or industries, there is still a limited number of studies that would comprehensively examine efficiencies of employment of entire economies, including their cross-sectoral relations. Moreover, their discussion does not exactly imply regarding how to improve the efficiency of sectors, organizations or other public and private entities as well as do not provide methodological guidance on how to achieve this in the context of SD.

Summing up the first step of the literature review, it is worth stressing that as all dimensions of human existence are interrelated, higher productivity, both within sectors and by shifting to sectors that generate higher value-added and yield benefits to workers, employers and the economy, is necessary to achieve sustained economic growth (World Development Indicators 2016). This idea led us to assess sectoral results for SD within EU countries.

Secondly, DEA has proven to be a useful tool for assessing productivity of entities which is of vital practical importance in managerial decision making (Park 2010). DEA is also widely considered as a technique to examine the level of efficiency and thus sustainability. Therefore, it is very important to carefully use DEA for guiding large policy issues such as an employment policy. This study may help to solve problems of low labour productivity by linking it with political efforts.

DEA is a method that allows relative technical efficiencies evaluation of a set of comparable entities, which use multiple inputs to produce multiple outputs. It uses linear programming (LP) techniques, in which a measurement of efficiency is defined as a ratio between the weighted sum of outputs and inputs. Analysed entities are usually called decision-making units (DMUs) (Poveda 2011). All DMUs are homogeneous, which means that they have the same types of inputs and outputs, and are independent, i.e. no constraint binds input and output levels of one DMU with inputs and outputs of another (Castelli et al. 2010). DMUs efficiency is measured in terms of radial or non-radial distance from the best unit on the production frontier, which is defined by the production function of the efficient units (Jain and Natarajan 2015).

SD, which seeks to harmonise social, economic and environmental capital, has become a reference point for rational policies, including policies on the allocation of labour resources. The task of improving the efficiency of resource allocation requires, firstly, the identification of the current state, secondly the diagnosis of the required shifts in employment, and finally the implementation of the suggested changes. The presented results may support the two first elements by providing general conclusions about the economic, social and environmental efficiency of EU employment, as well as suggestions for a shift in employment. It is worth stressing that even though there are some scientific analyses focused on the economic results of structural changes in employment, this paper concerns some neglected areas of compound SD goals. It is focused on specifying sectoral employment relation to multidimensional development and thus goes further than typical studies of aggregate productivity and growth.

Conclusions may be drawn regarding three areas:

Considering the first area, the main findings prove that a universal pattern of structural changes specified in the three-sectoral theory is favourable to SD in the EU. It was made clear that efficiency of employment is at its highest in the service sector and at its lowest in agriculture. This conclusion holds true for the compound process of SD; however, some special features describe each dimension. While employment in non-market services is found to be the most favourable to economy and society, market services are the best for the environment. Agricultural employment is the least efficient considering social and environmental gains, but it is outplaced by industry in the economic dimension. Nevertheless, the paper sheds light on some arguments in a discussion about de- and reindustrialisation. The detailed indices may differ concerning the set of aims adopted by policymakers, however, in all cases, growing share of service employment is both an inevitable objective trend and a signal of maturity of economy developing into sustainability. Thus, a structural policy cannot neglect servicisation process and it should rather direct it into a more inclusive, clean and knowledge-intensive one instead of set limitations with industrialisation slogans.

Comparisons among the three dimensions of development lead to the conclusion that the sphere that requires the most attention is the environmental one. It was shown to have the lowest efficiency and highest variability among the countries examined, especially between the “old” and “new” EU member states. Considering the strong external effects of the environmental sphere, differentiation leads to a decrease in the general living standards across the EU. Thus, common efforts are especially required to protect the natural environment.

Conversely, the highest level of employment efficiency was observed in the social sphere, where the examined 25 EU countries were also the least varied. This supports the general statement that employment, as an alternative to inactivity and unemployment, is favourable for social goals and may constitute the main instrument in social policy. The fact sets a huge challenge for an educational system, which must be able to prepare young people to successfully enter a labour market as well as for policy aimed at increasing the inclusivity of economy, especially in a face of growing flow of migrants. Moreover, additional challenge is specified by a necessity to increase sectoral mobility of workers, that is especially important for elderly employees.

Even though a high level of employment may induce gains in all dimensions of SD, the structural allocation of labour may differ in its importance. The results of this research suggest that large adjustments in the structure of employment are required in order to improve social efficiency, followed by environmental and economic efficiency. At first glance, this observation seems to be at odds with the observation that the highest employment efficiency is observed in the social sphere. As a high level of employment is generally better for the equal distribution of income, resulting in high efficiency, it is possible to improve it by adjusting the structure of employment. The fact once again stresses a necessity of de-agrarisation of employment and of a growth of welfare services. Additionally, the environmental efficiency of employment is low and it is difficult to improve it by making structural shifts. It proves strong externalities connected with environmental issues in any sector and allows for suggestion for wide horizontal rules in its protection.

Considering the third area of findings it is possible to point out a strong divide between the two groups of EU member states—the “old” and the “new” ones. The obtained results confirmed that the employment shift towards structures created in the highly developed EU countries may contribute to the improvement of employment efficiency across the EU. The highly developed countries were characterised by a relatively low level of employment in agriculture and achieved higher efficiency. A relatively low level of efficiency, in particular of the “new” EU member states, suggested a need for further reallocation of the labour force, i.e. the employment shift from less efficient sectors, primarily agriculture, to market and non-market services. This observation positively verifies the research hypothesis about a higher level of SD in structurally more mature economies. It also justifies the need for cohesion policy in the EU that supports the least developed countries as all dimensions of SD have some external effects for the whole European community.

The countries of the former Eastern Bloc are most in need of making employment shifts. In this group, Romania, Bulgaria, and the Baltic republics experience the strongest losses in efficiency. Apart from a low level of employment efficiency, the “new” countries face other problems, such as social stratification, income inequalities, low economic activity and early abandonment of the labour market. A significant share of the workforce in these countries was or is employed in the least productive sector, i.e. in agriculture. Their limited mobility, relatively poor health, lack of competencies required in a diversified labour market, or common reluctance to change jobs are great constraints in the flow of human resources. The solution to this situation, which facilitates the free movement of labour resources from low to highly efficient sectors, is to increase the knowledge, skills, and competencies of people working in the primary sectors. Thus, it seems to be necessary to cover such groups of people by a lifelong learning programs that are focused on increasing their intersectoral mobility. Moreover, it becomes more and more important to modernise a system of education to prepare young people for challenges of a knowledge-based service economy. Another political issue in this “new” member states is specified by a migration issue, which, in opposition to problems with a massive inflow of refugees to the West and North EU countries, is specified by net outflow of human resources and a “brain drain” problem. In a long-term it may be limited only by enhancing general efficiency and levering development and this requires rational structural policy. It is necessary to create such workplaces that are able to engage especially the young people with high ambitions, who are often university graduates and are just starting their professional career. Such jobs can be found mainly in the knowledge-intensive services.

Another problem in the majority of Central and Eastern European countries is the high level of environmental pollution caused by the use of energy-intensive technologies of production and the excessive greenhouse gas emissions associated primarily with the use of hard coal as a primary source of energy. This pollution is generated mainly in economies with a substantial share of agricultural employment. Suggested changes in the employment structure of these economies, relying on the concentration of labour resources in high-tech, knowledge-intensive sectors would not only contribute to an increase in efficiency but would significantly reduce pressure on the environment.

The research findings show that there exist benchmark economies that perform in the most efficient way while transforming employment into SD. Although each economy may adopt some unique pattern of development, “old” EU member states, such as Belgium, the Netherlands or Ireland, most often constitute the benchmarks for the catching-up economies. A high level of service employment and a de-agrarian trend are typical features of high-performing states. Moreover, the role of non-market services in these economies cannot be over-appreciated, pointing at the importance of welfare state solutions.

Summing up, this study may be useful to governmental institutions of individual EU countries in terms of their employment policy. Suggested changes in the employment structure can be used by decision-makers to improve employment efficiency. Research results may be useful not only for decision-makers from the EU countries, but also for those from the EU candidate countries, which already before the accession or when preparing for it (e.g. the Balkan states, Ukraine or Turkey) can make a diagnosis and analysis of employment efficiency and on this basis, develop strategies how to increase it.

The research method proposed in the paper can also be used in the diagnosis of the efficiency of areas such as environmental protection, education, healthcare or efficiency of privatisation and restructuring processes. Finally, the use of DEA provides a methodological linkage among economic, environmental and social areas of SD, so enhancing the practicality in mitigating their problems. This study provides guidelines for future research works on DEA on economic, social and environment issues. The sectoral results may still not be enough to formulate suggestions toward implementing reforms that lead to SD and the branch or even enterprise level may help to provide better guidance.