Chapter 3. What kinds of skills give countries a global advantage?1

The chapter analyses how different types of skills relate to export performance and participation in global value chains (GVCs) and investigates how skills characteristics shape countries’ comparative advantages in GVCs. To investigate the links between skills and GVCs, this chapter uses a new set of empirical analyses based on the Survey of Adult Skills and the Trade in Value Added (TiVA) database. It puts forward two major skills characteristics that shape countries’ comparative advantages in GVCs: the skills mix of the population and the role of pools of workers performing at the expected level. The chapter also indicates which industries countries could specialise in, given their skills sets, and what countries would need to do to specialise in technologically advanced industries.

While skills are believed to make a crucial contribution to performance in global value chains (GVCs), there is little evidence about how skills actually affect such performance. Trade experts consider that skills play an important role in countries’ trade performance and specialisation, and that a skilled workforce is a source of comparative advantage that enables countries to specialise in high-skilled segments of exports. The Heckscher-Ohlin model, a pillar of international trade theory, identifies skills as one of the factors that have a strong direct effect on countries’ industry specialisation and international integration. Studies that have tried to assess these links empirically share two types of limitation, however, both due to lack of data: skills are most often approximated by educational attainment and the emergence of this new pattern of trade, GVCs, is not taken into account.

Many OECD countries used to enjoy a comparative advantage because the education level of their population was higher, but this advantage is vanishing as tertiary education expands in many developing and emerging economies. Countries are increasingly competing not only through the education level of their population but also through the quality of skills, their effective use and efficient allocation of skills to industries. Chapter 2 showed that a broad range of skills can help countries to realise the benefits of GVCs. This chapter goes deeper by investigating which types of skills are important for participation and performance in GVCs. It also considers which characteristics of skills can shape specialisation in GVCs, and shows how these characteristics have to correspond to industries’ requirements for these industries to be able to export more.

To investigate the links between skills and GVCs, this chapter uses a new set of empirical analyses based on the Survey of Adult Skills, a product of the OECD Programme for International Assessment of Adult Competencies (PIAAC), and the Trade in Value Added (TiVA) database. In particular, this chapter:

Builds a set of new skills indicators based on the Survey of Adult Skills to characterise not only the cognitive skills of workers in each country, but also skills linked more closely to social and emotional aspects of jobs, which are particularly valued by employers. These skills indicators can also be used to describe the skills requirements of each industry.
Analyses how different types of skills relate to export performance and participation in GVCs, by investigating how countries’ skills in each industry are linked to their exports and GVC activities with various trade partners in the same industry. The chapter also examines how industries differ in their skills requirements.
Investigates how skills characteristics shape countries’ comparative advantages in GVCs. The chapter puts forward two major skills characteristics:
- The skills mix: Each individual needs to have high levels of several types of skills, rather than specialising in only one skill.
- Pools of workers: Most technologically advanced industries require pools of workers with reliable skills; such pools emerge in countries where individuals have the skills that would be expected given their various characteristics, including education level.
Explains the need to ensure that countries’ skills characteristics match industries’ skills requirements. The chapter indicates which industries countries could specialise in, given their skills sets, and what countries would need to do to specialise in technologically advanced industries.

The main findings in this chapter include:

When their workers have higher cognitive skills and stronger readiness to learn, as measured by the Survey of Adult Skills, countries tend to add more value to exports and participate more in GVCs.
Industries differ in their needs for skills in information and communications technologies (ICT); science, technology, engineering and mathematics (STEM); managing and communication; marketing and accounting; and self-organisation. In most industries, however – especially high-tech manufacturing and complex business services – workers perform several types of tasks requiring not only cognitive skills but also social and emotional skills. This means education systems need to develop a broad set of skills.
Skills policies can shape countries’ specialisation and give them a comparative advantage in GVCs, for instance by ensuring a better alignment of countries’ skills characteristics with the skills requirements of high-tech manufacturing and complex business service industries. By the same token, policies that favour a specific industry can lead to misallocation of skills and lower countries’ comparative advantage in other industries, generating costs for the economy.
Workers need to have a mix of skills to perform in an internationally competitive environment. Strong literacy or numeracy skills are not enough; workers also need strong problem-solving skills for technology-rich environments. Differences in the skills mix can lead to up to 60% differences in exports in one industry relative to another between two countries.
To be able to specialise in technologically advanced industries, a country’s population should, on average, have a level of the main skill required by an industry that is higher than that of other skills and higher than in other countries, and those with the higher level of the main skill should have the right mix of skills. Countries with the stronger alignment of the mix of skills with these industries’ skills requirements are Canada, Estonia, Israel, Korea and Sweden.
Countries need pools of workers with qualifications that reliably reflect what they can do (“reliable workers”) to be able to export more than other countries in high-tech manufacturing and complex business service industries, which require all workers to perform at the expected level. Pools of reliable workers emerge when individuals with similar characteristics (including education attainment) tend to share similar skills, such as in Japan, which can export (in value added terms) much more than Chile in high-tech manufacturing and complex business service industries relative to other industries. Japan, the Czech Republic, the Netherlands, and the Slovak Republic show a small dispersion of the skills of individuals with similar characteristics, helping them to provide pools of reliable workers.
Most OECD countries have gained comparative advantages in services and high-tech manufacturing industries. To maintain this specialisation, or specialise in other technologically sophisticated industries, countries have to ensure that overall, workers’ skills are strongly aligned with industries’ skills requirements. Countries whose skills characteristics are best aligned with technologically advanced industries’ requirements include the Czech Republic, Estonia, Japan, Korea and New Zealand. Australia, Ireland, the United Kingdom and the United States need to better align their skills characteristics with industries’ skills requirements to maintain or deepen specialisation in these industries.

Skills for economic performance

A taxonomy of skills

Thanks to the pioneering work of James Heckman, individuals’ skills – in all their diversity – are now recognised as fundamental determinants of economic and social success. Cognitive skills involve conscious intellectual effort and include long- and short-term memory, auditory and visual processing, processing speed, and logic and reasoning. Non-cognitive skills, also known as soft skills, social and emotional skills or personality traits, involve the intellect in a more indirect and less conscious fashion than cognitive skills, and relate to individuals’ personality, temperament, attitudes, integrity and personal interaction. Several analyses stress the importance of both cognitive and non-cognitive skills for occupational attainment and performance on the job (e.g. Heckman, Stixrud, and Urzua, 2006; Kautz et al., 2014).

General cognitive skills, which partly reflect the ability to learn, help predict the occupational level workers attain, their job performance and their ability to benefit from training (e.g. Schmidt, 2002; Schmidt and Hunter, 2004). There is strong empirical evidence that cognitive skills, rather than the level of schooling reached, influence individual earnings, the distribution of income and more generally economic growth (Hanushek and Woessmann, 2008).

Numeracy and mathematical skills are directly conducive to business success, particularly in technologically advanced industries (Hoyles et al., 2002). Many of the fastest-growing occupations and emerging industries require numeracy, knowledge of scientific and mathematical principles, as well as the ability to generate, understand and analyse empirical data and solve complex problems (UKCES, 2011). These skills make technological breakthroughs possible.

ICT skills play a key role in improving companies’ performance. Firms with high ICT capabilities tend to outperform comparable firms in the same industry on a sustained basis (Bharadwaj, 2000; Santhanam and Hartono, 2003). ICT investment seems to pay off for some companies but not others, because of organisational learning and ICT competencies in particular (Tippins and Ravipreet, 2003).

Along with cognitive skills, a wide range of personality traits matter for economic performance (Heckman and Rubinstein, 2001). Some authors argue that for many outcomes, these skills are just as important as cognitive skills, or even more so (Kautz et al., 2014). Many researchers group personality measures under five key factors: extraversion, agreeableness, conscientiousness, emotional stability, and openness to experience (Goldberg, 1990). Agreeableness includes skills like empathy, perspective taking, co‐operation, and competitiveness. Conscientiousness includes grit, perseverance, delay of gratification, impulse control, achievement striving, ambition, and work ethic. Emotional stability includes self-evaluation and self-esteem, self-efficacy and optimism. Many of these are a mix of traits that individuals are born with and abilities that can be learnt and improved over time.

On the job, some specific skills such as communication, management, self-organisation and problem solving are highly valued by employers and contribute to firm performance (Hitt, Ireland and Hoskisson, 2012; Bloom and Van Reenen, 2010; Bloom et al., 2012; Ichniowski, Shaw and Prennushi, 1997). These skills combine cognitive skills and personality traits.

Overall, skills that matter for job performance form a continuum, from skills that are mostly cognitive to skills mostly linked to personality traits, with skills in between that combine both (Figure 3.1, first part). In addition, physical skills are crucial in several sectors, such as construction, health and well-being, and the arts.

Building skills indicators from the Survey of Adult Skills

The Survey of Adult Skills provides a broad range of information about adults’ skills and the tasks they perform. This information can be used to measure some of the skills that researchers have identified as important for job performance and firm performance. The survey assesses three domains of cognitive skills (numeracy, literacy and problem solving in technology-rich environment) through administered tests. In addition, a background questionnaire asks how often individuals perform tasks including reading, writing, numeracy, ICT and problem solving, partially matching the set of cognitive skills assessed through tests. The survey also gathers information on how often other types of tasks are performed, such as those involving management, communication, organisation and planning, and physical work. Finally, the survey provides information on attitudes towards learning, trust, health and other issues.

Alongside the three cognitive skills assessed in the Survey of Adult Skills, the large set of information related to the skills of individuals has been summarised in six task-based skills indicators through a statistical method (Box 3.1): ICT skills, readiness to learn, management and communication skills, self-organisation skills, marketing and accounting skills, and STEM skills (Figure 3.1, second part). Since most of the task-based skills indicators2 are based on information about how often tasks are performed, they do not directly capture the skills possessed by workers.

Box 3.1. Developing a taxonomy of performance-relevant skills based on the Survey of Adult Skills

In a first step, a set of skills indicators was developed following a normative approach (Grundke et al., forthcoming a). Based on an extensive literature survey on the determinants of performance on the job and firm performance, 17 skills indicators that could be grouped into five categories were constructed, initially using the Survey of Adult Skills. The idea behind this categorisation was that skills should be seen as a continuum, with some mainly cognitive, others close to personality traits and a large group combining the two aspects.

The normative approach leads to skills indicators that can be easily interpreted, but it does not ensure that they are statistically relevant in terms of the covariance structure of the question items from the Survey of Adult Skills. In a second step, a set of new skills indicators were derived from an exploratory factor analysis, relying thus on a data-grounded method. An exploratory factor analysis assumes the existence of a certain number of unobserved latent variables, called factors (the new skills indicators), whose joined variation explains the correlation pattern of a larger set of observed variables. Each factor is a weighted combination of the observed variables, whereby the weights on the observed variables are called factor loadings. The number of factors is a parameter of the model and needs to be chosen carefully using certain criteria established in the literature (Conti et al. 2014; Costello and Osborne 2005).

One of the main drawbacks of the classical exploratory factor analysis is that observed variables may be associated with all factors, making the factors difficult to interpret. This issue can be addressed by following a three-step procedure (as in Costello and Osborne, 2005), which ensures that each observed item contributes to no more than one single factor. In a first step, factors are rotated to form groups of items loading on certain factors. In a second step, items that load on at least two factors with similar loadings (so-called double loadings) are dropped. Finally, in a third step, only items with loadings above a threshold of 0.45 are assigned to a certain factor.

As a result of the factor analysis, 33 items – variables from the background questionnaire of the Survey of Adult Skills – were retained from an initial set of 57 items. They were grouped into six factors that can be interpreted on the basis of the normative typology as follows: ICT skills, readiness to learn, management and communication skills, self-organisation skills, marketing and selling skills, and STEM skills – which are called task-based skills in this chapter, as opposed to skills that are assessed through test.

ICT skills: This consists of ten items with very high positive loadings and one item with a negative loading. The items with positive loadings all describe tasks associated with ICT use, from reading and writing emails to using word-processing or spreadsheet software, or a programming language. The factor is strongly associated with office jobs, as indicated by negative loading on “physical activities”.

Readiness to learn consists exclusively of items designed in the Survey of Adult Skills to measure this dimension, e.g. “Relate new ideas into real life” or “Like learning new things”.

Management and communication skills: This gathers a more diverse set of items, from “teaching people” to “planning others’ activities”. All these activities involve communicating with and managing other people, whether they are co-workers or not.

Self-organisational skills, like readiness to learn, consist exclusively of items designed in the Survey of Adult Skills to measure this dimension. It includes items such as “Work flexibility – Speed of work” or “Work flexibility – Sequence of tasks”.

Marketing and accounting skills is a newly constructed indicator that does not correspond to any indicator in the normative typology. “Reading financial statements”, “calculating costs or budgets” and “selling products or services” are associated with this factor, as well as “using a calculator”. Although the last item also loads on “ICT contents” and “STEM contents” (with loadings close to 0.25), it seems that calculators are mainly used for marketing and accounting purposes.

STEM skills: This factor has not been present in the normative typology. Like Marketing and Accounting Skills, it involves numeric tasks such as “Use Simple algebra or formulas” or “Use advanced math or statistics”, but they are more complex and less specific than those loading on the previous factor. This factor is interpreted broadly as skills necessary for Science, Technology, Engineering and Mathematics.

Each task-based skills indicator has a score ranging from 0 to 1. A higher score is associated with a higher frequency of performing these tasks on the job.

Source: Conti, G. et al. (2014), “Bayesian exploratory factor analysis”, Journal of Econometrics.

Costello, A.B and J.W. Osborne (2005), “Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis”, Practical Assessment, Research & Evaluation.

Grundke, R. et al. (forthcoming a), “Skills and global value chains: Characterisation and evidence”, OECD Science, Technology and Industry Working Papers.

Some relevant skills dimensions cannot be measured, while others are measured only imperfectly, because there is always a gap between conceptualising skills and measuring them. When subjected to measurement, personality traits tend to include a component of cognitive aspects and cognitive skills also depend on individuals’ personality traits. In addition, while the Survey of Adult Skills offers a wealth of information, it was not conceived to measure all the various skills needed at work. The set of items used in developing the skills indicators is constrained by the list of items available in this survey.

In particular, while many of the personality traits affect work outcomes, openness to experience is the only trait that can be represented by an indicator stemming from the Survey of Adult Skills, the readiness to learn indicator. Openness to experience can improve firm performance by encouraging workers to undertake training and adapt when they face unfamiliar environments. It also appears important for complex jobs that involve autonomy and require unconventional thinking and the adoption of new behaviours and ideas to achieve high job performance (Mohan and Mulla, 2013).

The different types of skills indicators available through the Survey of Adult Skills provide for more precise measures of skills than educational attainment, which is frequently used as a proxy for skills, including in most of the empirical literature on skills and trade. Educational attainment contributes to cognitive and other skills but disguises differences in the quality of countries’ education systems. It does not account for differences in the way skills are developed on the job or for the large range of skills that could influence countries’ performance within GVCs.

Skills patterns across countries and industries

Countries differ in their workers’ skills sets. For the three domains of assessed cognitive skills, Japan and Finland have the most proficient workers, while workers in Greece, Turkey, Chile and Italy have the lowest scores, on average, among OECD countries (Figure 3.2). The picture is much more mixed in terms of task-based skills (Figure 3.3). The top performers in cognitive skills tend to rank well in terms of task-based skills, while some countries like the United States, whose workers have average or below average cognitive skills, rank high on these indicators.

The variation between countries depends on the skills indicator (Figures 3.2 and 3.3). Countries tend to differ the most in terms of cognitive skills assessed through tests (literacy, numeracy and problem solving in technology-rich environments), and the least for marketing and accounting, and STEM skills, two task-based skills. Large cross-country differences also emerge for readiness to learn, with Japan and Korea appearing at the bottom, which reveals the importance of cultural norms. However, given that this indicator is based on self-reported information, it is difficult to know the extent to which differences between countries emerge from real differences in attitudes towards learning or in the way questions are answered.

In terms of countries’ performance in GVCs, it is not just countries’ levels of skills that are important, but also their allocation of skills across industries, which reflects their capacities to specialise and perform well in some activities and industries. On the one hand, industries requiring higher skilled workers and those in which countries tend to perform well because of other reasons than skills (e.g. historical reasons or the availability of some type of capital specific to the industry) should attract more skilled workers, but this depends on countries’ overall skills level and on the efficiency of the allocation process. On the other hand, industries in which countries perform well make greater efforts to enhance their workers’ skills through training.

For all types of skills, workers in business services have on average higher skills than workers in other industries (Figures 3.4 and 3.5). Within a given industry, the heterogeneity of workers’ skills across countries varies by type of skills. In most industries, large differences exist between countries in terms of cognitive skills assessed through tests and readiness to learn reflecting the dispersion of these skills across countries. Heterogeneity is much lower for the other task-based skills, revealing that industry characteristics play a bigger role than country specificities in determining the allocation of these skills (Figure 3.5).

Large variations emerge among industries in the use of ICT skills, with business services making the highest use. Among the high-skilled industries (such as business services), some countries make a much lower use of ICT skills than the top country performers. Insufficient ICT skills for some countries in some industries raise concerns, as this can dampen their capacity to grow and upgrade activities in the current technology-rich environment.

As with other task-based skills, marketing and accounting skills exhibit large variations between industries, yet minimal variations among countries. Two patterns emerge when comparing different industries. First, business services industries (with the exception of wholesale and retail trade), along with construction and transportation, display larger cross-country gaps, which may reflect the specialisation of some countries in marketing and accounting tasks. Second, not only are the gaps among these services industries larger, but their medians are also much higher than those in the manufacturing sector. This result is in line with the nature of occupations within these industries, involving intense interaction with customers, as well as the trend among manufacturing companies to outsource the marketing and distribution activities of their products to service companies, which are better trained to perform these tasks.

The need for a diversity of skills

The varying role of different types of skills for export performance and participation in global value chains

Workers’ skills are generally considered crucial for countries’ participation in GVCs and export performance, but little is known about which skills are most relevant for each type of export and each kind of participation in GVCs. Some skills, particularly cognitive abilities, could contribute directly to value creation in firms (Barney, 1991; Wright, McMahan and McWilliams, 1994) and subsequently add more value to exports and to intermediates exported for use in third countries’ exports (forward linkages or participation). Other skills may encourage offshoring of activities and the use of intermediates from abroad (backward linkages or participation).

Skills can matter for integration into GVCs not only in their diversity but also in the way workers’ proficiency varies. A large volume of international activity takes place between developed countries with similar average skills and technologies. The abilities of professionals, production workers and other workers differ among otherwise similar countries, which might explain why these countries benefit from trade.

As countries and industries differ in terms of their levels of skills and their participation and performance in GVCs, it is possible to shed light on the relationship between the two by assessing the links between countries’ skills levels by industry, on one hand, and their exports and GVC activities with various trade partners in the same industry on the other hand (Box 3.2).

Box 3.2. The empirical links between different types of skills and performance in GVCs

The discussion in this section is based on work testing how various types of skills are related to trade and participation in GVCs (Grundke et al., forthcoming a). For this purpose, several indicators of exports and participation in GVCs are linked to various skills indicators one by one, measured in terms of average skills level, skills dispersion, and the median, upper and lower parts of the skills distribution. The skills indicators include the three assessed cognitive skills from the Survey of Adult Skills (literacy, numeracy and problem solving in technology-rich environments), and the six task-based skills indicators resulting from the factor analysis, as described in Box 3.1 (ICT skills, readiness to learn, management and communication skills, self-organisational skills, marketing and accounting skills, STEM skills). They are all country-industry specific.

All indicators of trade and participation in GVCs stem from the Trade in Value Added Database and are used in the analysis at the bilateral industry level and in log form. Exports are considered in gross and value added terms. Three indicators of participation in GVCs are considered: domestic value added embodied in foreign final demand for forward participation (or linkages) in terms of final demand; foreign value added embodied in domestic final demand for backward participation in terms of final demand; foreign value added in exports for backward participation in terms of exports.

The model also includes a series of independent country-industry variables – physical capital intensity, human capital intensity (measured by educational attainment) and expenditure for research and development –to reflect how well countries are able to meet the requirements of industries along various technological dimensions. Barriers to trade are also included, as well as fixed effects to account for country, partner country and industry characteristics.

Country-industry groups that have fewer than 25 observations in the Survey of Adult Skills are removed to reduce measurement error. Standard errors are clustered at the importer-exporter pair level.

Source: Grundke, R. et al. (forthcoming a), “Skills and global value chains: Characterisation and evidence”, OECD Science, Technology and Industry Working Papers.

Such analysis confirms that cognitive skills and personality traits matter for exports, in gross and value added terms, and for participation in GVCs (Figure 3.6). Literacy, numeracy, problem solving in technology-rich environments and readiness to learn all tend to be stronger where exports are stronger, even more so when exports are expressed in value added terms, with cognitive skills having the strongest links. These skills are also likely to be higher where participation in GVCs is stronger, through both backward and forward linkages. These results support the idea that knowledge and learning play a fundamental role in international integration as workers need them to apprehend, share, and assimilate new knowledge in order for countries to participate and grow in evolving markets.

The cognitive skills of different groups of workers – medium, low and high skilled – matter in different ways for exports in value added terms and participation in GVCs (Figures 3.7 and 3.8). For literacy, the skills of the most proficient matter most. By contrast, numeracy skills seem to be important across the industry (i.e., at the median of the distribution). This suggests that numeracy is needed not only for innovation and value creation for exports (reflected in forward linkages and exports in value added terms) but also for integrating foreign value added in the domestic production process (backward participation). For problem-solving skills in technology-rich environments, both low-skilled workers (the lower 10th percentile) and high-skilled ones (90th percentile) need to be proficient in their categories to bring domestic value added to international markets (forward participation) as well as to process inputs at the assembly line (backward participation).

Readiness to learn3 across various types of workers is also likely to be higher where performance in GVCs is stronger. Those most ready to learn tend to have the strongest effect on export activities and participation in GVCs.

In an increasingly competitive and technology-based global economy, it is not surprising to see a positive link between ICT skills and exports, especially when expressed in value added terms (Figure 3.6). The negative link between ICT skills and STEM skills, on one hand, and participation in GVCs, on the other hand, is more ambiguous. A possible explanation is that certain tasks in industries and occupations intensive in these skills are typically more difficult to offshore.4

Many studies have underlined the relevance to firms’ performance of skills combining cognitive and personality trait aspects such as management, communication, marketing, and self-organisation. However, some of these skills appear to be only weakly related to performance and participation in GVCs.

Managing and communication skills matter for exports in value added terms, and for international integration in the case of forward linkages (Figure 3.6). Communication and interaction skills can develop complementarities among workers in production, facilitate gains from specialisation, and encourage gains from knowledge transfer, which in turn would benefit forward and backward participation in GVCs alike. Combined with strategic management abilities, communication and interaction skills can generate sustainable advantages and enhance competitiveness in global markets.

Strong marketing and accounting skills, as well as self-organisation skills, do not appear to be important for exports and integration in GVCs5 (Figure 3.6). Self-organisation skills may be negatively linked with backward participation because these skills are most frequently used by managers, who are less concerned by offshoring. Although these skills may not by themselves be strongly linked to performance in GVCs, they could have an impact when considered jointly with other skills and firm capabilities. Such counterintuitive results might also stem from the fact these indicators are based on questions about the frequency of the performance of related tasks, which imperfectly approximate the skills workers have.6

Some levels of task-based skills of groups of workers relate negatively to GVC indicators while others relate positively (Figures 3.7 and 3.8). For instance, the higher the ICT skills and managing and communication skills of the bottom percentile, the less industries use foreign intermediates in their exports, but the higher the skills of the top percentile, the more so industries use foreign intermediates. This suggests that the use of foreign inputs could be a substitute for low ICT skills and managing and communication skills, but needs to be complemented by high levels of these skills for top performers. In contrast, the higher the STEM skills of the lower part of the distribution, the more industries export, suggesting that these skills are important for low-skilled workers in an international environment. Hence, some skills need to be possessed by all workers while others are particularly important for low or top performers.

The results presented here suggest that policies boosting certain workers’ levels of different skills could play an important role in raising countries’ integration into GVCs. However, these results do not show a causal link from skills supplied at the country-industry level to the performance of industries in GVCs. Since workers are mobile between industries, changes in the economic performance and participation in GVCs of these industries might strongly influence the allocation of workers among industries within a given country and thus the supply of skills – especially of task-based skills – at the country-industry level.

How industries differ in their skills requirements

If they are taken at an industry level, the task-based skills indicators computed from the Survey of Adult Skills (ICT, STEM, managing and communicating, marketing and accounting, and self-organisation; see Box 3.1) reflect the extent to which an industry involves more of these tasks and hence requires more of the necessary skills. According to these indicators, most industries are intensive in a broad range of tasks, suggesting that they require workers with a broad range of skills (Figure 3.9).

As might be expected, more complex business services and high-tech manufacturing are more task-intensive than less complex business services and low-tech manufacturing. However, even less complex business services and low-tech manufacturing use a broad range of skills, including management and communication skills and self-organisation skills. High-tech manufacturing is more intensive in STEM tasks but also frequently involves tasks requiring “soft” skills such as managing and communicating. Overall, it is difficult to characterise industries by the performance of one specific task. Several different types of industries are intensive in ICT and STEM tasks.

To gain a clearer picture of industries’ skills needs, by gauging the proportions in which different industries need specific skills, the same task-based skills indicators can be used to compare the frequency of one specific task with the frequency of another (Figure 3.10). Many high-tech manufacturing sectors show higher intensity in self-organisation tasks than in management and communication tasks, or marketing and accounting tasks.7 Business services tend to be more intensive in marketing and accounting tasks than in communication tasks, with some complex business services also intensive in self-organisation tasks. These results support the idea that a broad set of skills is required by industries and that industries can be characterised by their relative intensity in the performance of these tasks.

Understanding countries’ specialisation in GVCs

Countries differ in their skills endowments while industries vary in their skills requirements. This section examines how the interaction between countries’ skills characteristics and industries’ requirements partly explains why countries perform better in some industries within GVCs, as measured by the extent to which they export more in value added terms in one industry rather than in another one. While the focus is put on skills, other factors are taken into account, such as physical capital and trade costs that also influence specialisation in GVCs.

Having the right mix of skills

Employers expect workers to have a mix of skills, and firms’ performance depends on this diversity of skills. Within GVCs, workers may have to perform technical tasks but also to communicate with foreign colleagues, requiring them to have technical and communication skills. The extent to which workers have the right mix of various skills influences countries’ performance within GVCs. For instance, a survey on India found that only one in four graduates of engineering schools were employable, as most were deficient in at least one of the required skills – technical skills, fluency in English, the ability to work in a team or to deliver basic oral presentations (Ohnsorge and Treffler, 20078). Concerns about the skills of Indian graduates arose in a context of increasing offshoring from developed economies and competition to attract foreign investments. Likewise, studies based on employer surveys in Europe put forward the role of a combination of social and emotional skills, technical skills and cognitive skills for firms’ performance (Humburg, van der Velden and Verhagen, 2013). The lack of interpersonal skills can create a strong barrier to employment, especially for low-skilled jobs (Heckman and Kautz, 2013). These studies show that it is important for each worker to have the right mix of skills, rather than for firms to have a set of workers with each specialised in one skill.

While there is a broad consensus that having the right mix of skills is important for employability and firms’ performance, very few studies try to characterise countries’ endowment in terms of combinations of skills or to assess the impact of skills mixes on performance in GVCs.

To understand how the skills mix can shape countries’ specialisation in GVCs and how this can be assessed empirically, it is useful to consider two skills (for instance numeracy and literacy skills, or quantitative and communication skills). Workers can be characterised by the absolute levels in these two skills, their absolute skills advantage, and whether they are better in one skill than another – the relative skills advantage. Industries differ according to their relative skills requirements, as illustrated in Figure 3.10. The allocation of workers across industries is not determined by how much of each type of skills a worker has, but by the ratio of one skill to the other (workers’ relative skills advantage). For example, a worker with a high ratio of numeracy to literacy skills works in industries more intensive in quantitative tasks. The absolute value of skills does not influence the sorting of workers across industries, but it does affect the productivity of workers, who in fact need both numeracy and literacy skills.

A country’s capacity to be internationally competitive in certain industries and to specialise in these industries depends on the country-wide correlation between the relative and absolute skills advantage across the population. In the above case of a relative skills advantage measured as a high ratio of numeracy to literacy skills, and of an absolute advantage measured as strong literacy and numeracy skills, a country with a high positive correlation between the two will specialise in quantitative-intensive industries. The best workers – those with an absolute advantage in both skills – will sort into the quantitative-intensive industry and thus increase the absolute productivity of this industry compared with countries where the correlation is lower. Conversely, the country with the lower correlation would specialise in industries requiring literacy skills such as those intensive in communication tasks, as its workers have an absolute advantage in communication-intensive industries relative to other countries (Box 3.3).

Box 3.3. The empirical link between countries’ skills mix and specialisation in GVCs

The discussions in the current section and the following one are based on OECD work testing the role that each country’s skills mix plays in its specialisation in certain industries in GVCs (Grundke et al., forthcoming b). The empirical specification is based on a theoretical model which assumes that workers are heterogeneous and endowed with a mix of two skills (e.g. quantitative and communication skills) (Ohnsorge and Treffler, 2007). Industries differ according to their skills requirements, while the marginal product of a specific skills mix differs across industries. The key parameter explaining countries’ specialisation in certain industries is the country-wide correlation between the relative and absolute skills advantage across the population. This correlation indicator, jointly with the relative skills endowment, explains countries’ comparative advantage in GVCs.

This publication uses the assessed skills of literacy, numeracy and problem solving in technology-rich environment from the Survey of Adult Skills to measure the country-wide skills mix. All three possible combinations of skills mixes (numeracy to literacy, problem solving to numeracy and problem solving to literacy) are tested. The relative intensity of industries in these three types of skills is computed by using the task-based skills indicators from the factor analysis at the industry level (Box 3.1), after establishing a correspondence between the assessed skills and the industry intensity in tasks that relate to these specific skills (Table 3.1). Based on the description of the cognitive assessment tests in the Technical Report of the Survey of Adult Skills, it is reasonable to assume that literacy skills are more in demand in industries intensive in management and communication tasks, as measured by the task-based skills indicator. Similarly, numeracy skills can be associated with the task-based skills indicator marketing and accounting, and problem solving in technology-rich environments with the task-based skills indicator self-organisation.

To test the importance of skills mixes for countries’ specialisation in GVCs, the empirical analysis explains exports in value-added terms in each industry of a country towards its trade partners by the country-specific correlation of relative and absolute advantage of workers (in two types of cognitive skills) in relation to the relative intensity of industries in two specific tasks (that correspond to the two cognitive skills). The other major explanatory variable is the country-industry interaction of the relative skills advantage with the relative task intensity of industries. Each specification is estimated for each possible combination of two assessed skills (numeracy to literacy, problem solving to numeracy and problem solving to literacy) with their respective corresponding relative industry task intensity (marketing and accounting to management and communication, self-organisation to marketing and accounting, and self-organisation to management and communication).

The empirical analysis uses the typical sectoral gravity model for bilateral trade flows that are used in the empirical literature on comparative advantage (Romalis 2004, Nunn 2007, Levshenko 2007, Chor 2010). The constructed bilateral industry-level dataset includes 23 exporting countries, 62 importing countries (including rest of world) and 34 TiVA industries. Exports in value-added terms are taken for the year 2011 from the TiVA 2015 database. All specifications include the final demand at the importer-industry level as an independent variable. Additional explanatory variables include traditional Heckscher-Ohlin country-industry measures of relative endowments of physical and human capital, bilateral trade costs variables from the CEPII GeoDist database (Mayer and Zignago, 2011) and fixed effects to account for exporter, importer and industry characteristics, as well as dummy variables that control for all omitted aggregated sector characteristics for the exporting and importing country (34 industries are aggregated into the three sectors resource extraction, manufacturing and utilities, and services). Robust standard errors are clustered at the exporter-importer level.

Sources: Chor, D. (2010), “Unpacking sources of comparative advantage: A quantitative approach”, Journal of International Economics.

Grundke, R. et al. (forthcoming b), “Having the right mix: The role of skills bundles for comparative advantage and industry performance in GVCs”, OECD Science, Technology and Industry Working Papers.

Levchenko, A.A. (2007), “Institutional quality and international trade”, Review of Economic Studies.

Nunn, N. (2007), “Relationship-specificity, incomplete contracts, and the pattern of trade”, The Quarterly Journal of Economics.

Mayer, T. and S. Zignago (2011), “Notes on CEPII’s distances measures: the GeoDist Database”, CEPII Working Paper 2011-25.

Ohnsorge, F. and D. Treffler (2007), “Sorting it out: International trade with heterogeneous workers”, Journal of Political Economy.

Romalis, J. (2004), “Factor proportions and the structure of commodity trade”, American Economic Review.

The Survey of Adult Skills gives information on the mix of skills of the population. As it includes assessment of proficiency in three skills – literacy, numeracy, and problem solving in technology-rich environments – it is possible to investigate how these various skills are correlated to each other.

Countries vary in their population’s skills mix (Figure 3.11). In a group of countries, including the United States, workers who are more proficient in numeracy than in literacy skills (with a relative skills advantage in numeracy) also have high literacy skills (an absolute skills advantage) while the reverse is true in other countries, such as the Czech Republic. This means that in the United States, the part of the population with a relative advantage in numeracy skills (high ratio of numeracy to literacy scores) is also the one with higher absolute scores in both cognitive skills, i.e. numeracy and literacy. In contrast, in the Czech Republic, workers with a relative advantage in numeracy skills have low absolute scores in both skills. In all countries, workers whose skills in problem solving in technology-rich environments are higher than either their numeracy or literacy skills lack high numeracy or literacy skills in absolute terms, but there are variations among countries.

To understand how workers are shared among industries, a correspondence between workers’ cognitive skills and industries’ skills requirement needs to be established (Box 3.3, Table 3.1). As cognitive skills measured in the Survey of Adults Skills capture a large set of abilities, they can be matched to industries’ skills requirements. The literacy dimension more broadly measures the ability to analyse complex social contexts and to deal with social interaction using a language, which is expected to be required in industries intensive in management and communication tasks. The numeracy dimension measures the ability to understand, use and communicate mathematical information, and is thereby expected to be needed in industries intensive in marketing and accounting tasks. Problem solving in technology-rich environments includes identifying problems, setting goals and being self-organised to find solutions. This type of skill is important in industries intensive in self-organisation tasks.

Table 3.1. Correspondence between industries’ task intensities and skills requirements
Industry task intensity	Correspondence in terms of cognitive skills requirement
Managing and Communication	Literacy
Marketing and Accounting	Numeracy
Self-Organisation	Problem solving in technology-rich environments
Source: OECD calculations based on the Survey of Adult Skills (PIAAC) (2012), www.oecd.org/skills/piaac/publicdataandanalysis.

Industries vary in their task intensities, and countries differ in their skills mix to provide for these industry-specific skills requirements; comparative advantage within GVCs stems from such country-industry matches. Estimates using the Survey of Adult Skills and TiVA database show that:

Countries with highly correlated relative skills advantage in numeracy as compared with literacy and absolute advantage in these skills can export more in industries that are more intensive in marketing and accounting tasks than management and communication tasks (Figure 3.12, Panel A). This is the case, for instance, for Australia, Ireland, Norway, Turkey and the United States, where workers with higher numeracy skills than literacy skills also have high literacy skills in absolute terms. Their skills mix gives these countries a comparative trade advantage in many business services industries, both complex (including finance and insurance, and research and development) and less complex (like wholesale and retail trade) (Figure 3.10).
Countries with highly correlated relative skills advantage in problem solving in technology-rich environments as compared to literacy skills, and absolute advantage in these skills, can export more in industries that are more intensive in self-organisation tasks than in management and communication tasks (Figure 3.12, Panel B). This is the case for the Czech Republic, Japan, Korea, Poland and Slovenia, where workers whose skills in problem solving in technology-rich environments are higher than their literacy skills also have high literacy skills in absolute terms. Their skills mix gives these countries a comparative trade advantage in many complex business services industries, including computer and related activities, and finance and insurance, as well as some high-technology manufacturing industries like chemicals and computer products (Figure 3.10).
Countries with highly correlated relative skills advantage in problem solving in technology-rich environments as compared to numeracy skills and absolute advantage in these skills can export more in industries that are more intensive in self-organisation tasks than marketing and accounting tasks (Figure 3.13, Panel C). This is the case, for instance, for the Czech Republic, Estonia, Japan, Poland and Slovenia, where workers whose skills in problem solving in technology-rich environments are higher than their numeracy skills also have high numeracy skills in absolute terms. Their skills mix gives these countries a comparative trade advantage in mostly business services, in chemicals high-tech manufacturing, as well as in low-tech industries like pulp and paper products (Figure 3.10).

The role of relative skills endowment

A country’s tendency to specialise in an industry depends not only on having the right mix of skills, but also on its relative skills endowment. Countries can capture larger shares of world production and trade in GVCs in industries that use more intensively their abundant skills. Studies that have looked at how educational attainment contributes to comparative advantage show that countries abundant in highly educated workers specialise in industries that use this factor intensively (Romalis, 2004). Such studies have used educational attainment as a proxy for skills, due to data limitations.

The Survey of Adult Skills makes it possible to investigate the role of relative skills endowment in different types of skills, rather than just classifying workers as high-skilled or low-skilled. If two skills are considered, numeracy and literacy, and two tasks, quantitative and communication ones, a country in which the population is more skilled in numeracy than in literacy can export more in industries that are more intensive in quantitative tasks than in communication ones. A country in which the population is less skilled in numeracy than in literacy can, vice versa, export more in communication-intensive industries. Using the Survey of Adult Skills, the relative skills endowment of a country can be measured by the average ratio of scores in terms of two skills, such as numeracy and literacy, and the average ratio of scores in problem solving in technology-rich environments to literacy and numeracy.

Compared with other countries, the population appears to be skilled in: 1) numeracy relative to literacy in Austria, Belgium and Denmark; 2) problem solving in technology-rich environments relative to literacy in Chile, Germany, Israel and Turkey;9 and 3) problem solving in technology-rich environments relative to numeracy in Chile, Turkey, the United Kingdom and the United States (Figure 3.13).

As with the correlation between relative and absolute skills advantages, the differences in the relative skills endowments of workers across countries generate comparative advantages within GVCs in some industries depending on their skills characteristics.

Countries where numeracy skills are higher than literacy skills can export more in value added terms in industries that are more intensive in marketing and accounting tasks than in management and communication tasks (Figure 3.14, Panel A).
Countries where skills involving problem solving in technology-rich environment are higher than literacy skills can export more in value added terms in industries that are more intensive in self-organisation tasks than in management and communication tasks (Figure 3.14, Panel B).
Countries where skills involving problem solving in technology-rich environment are higher than numeracy skills can export more in value added terms in industries that are more intensive in self-organisation tasks than marketing and accounting tasks (Figure 3.14, Panel C).

The relative endowment and the correlation between relative and absolute skills advantage contribute to countries’ specialisation in GVCs. To be able to specialise in a specific industry, the population should on average have a higher level of the main skill required by this industry compared with other skills than other countries and those with the higher level of the main skill should also have the other skills required by this industry. Except for the combination of literacy and numeracy skills, the comparative advantages in GVCs stemming from the relative skills of the population is smaller than those emerging from having the right mix of skills. These two determinants of specialisation can be combined in an overall skills mix effect (Box 3.5).

The role of countries’ skills dispersion: providing pools of reliable workers

Each country’s dispersion of skills influences what industry it specialises in, as well as its competitiveness patterns. Even if two countries have identical average skills endowments, they will trade with each other depending on the properties of their human capital dispersion (Asuyama, 2012; Bombardini, Gallipoli and Pupato, 2009; Bougheas and Riezman, 2007; Grossman, 2004 and 2013; Grossman and Magi, 2000). This publication is the first to investigate the role that skills dispersion plays in specialisation patterns within GVCs (Box 3.4).

Box 3.4. Analysing the effect of countries’ unobservable skills dispersion on specialisation in GVCs

The discussion in the current section is based on OECD work assessing the effect of countries’ skills dispersion on industry specialisation in GVCs (Grundke et al., forthcoming b). The empirical specification is based on a theoretical model that assumes workers are heterogeneous and production requires teams of workers (Bombardini, Gallipoli and Pupato, 2009 and 2012). Industries differ in the extent to which workers’ skills within the production team are complementary or substitutable. Some industries, especially those involving long sequences of tasks, require all workers to perform at the expected level, while others, in which skills are more easily substitutable, can cope with workers of low performance. The key parameter explaining countries’ specialisation in certain industries is the dispersion of skills after accounting for the observable skills determinants, hence the unobservable skills dispersion. According to the model, a country with a narrow dispersion of unobserved skills exports more in industries characterised by higher complementarity of workers’ skills in the production process than in industries with a lower skills complementarity.

Measures of countries’ unobserved skills dispersions and industries’ complementarity are based on information available in the Survey of Adult Skills. The unobserved skills are calculated by taking, for each worker, the difference between his/her literacy score and the estimated literacy score of a worker with similar characteristics in terms of education, age, gender, immigrant status and participation in adult education or training programmes 12 months before the survey date. The dispersion of these unobserved skills gives the skills dispersion that cannot be explained by countries’ characteristics. Industries’ degrees of complementarity are approximated by the average across countries for each industry of the task-based skills indicator management and communication, derived from the factor analysis (Box 3.1). This approach follows other studies that have used the O*NET database to approximate industries’ degree of skills complementarity (Bombardini, Gallipoli and Pupato, 2009 and 2012). The frequency of management and communication tasks reflects the importance of coordinating tasks to achieve a given level of output quality and thereby characterises industries’ degree of complementarity. Industries are ranked according to the complementarity index.

The empirical specification follows past studies (Bombardini, Gallipoli and Pupato, 2009 and 2012). The aim is to test the importance of unobservable skills dispersion for countries’ specialisation in GVCs. The empirical analysis explains exports (in value added terms) in each industry of a country towards its trade partners by the country-specific unobservable skills dispersion (for literacy skills) in relation to the industry degree of complementarity. The degree of complementarity is measured by the task-based skill indicator management and communication, as calculated at the industry level across all countries participating in the Survey of Adult Skills.

The empirical analysis follows the same approach and uses the same data as for the assessment of the role of the skills mix for specialisation in GVCs (Box 3.3). It uses the typical sectoral gravity model for bilateral trade flows. All specifications include final demand at the importer-industry level as an independent variable. Additional explanatory variables include traditional Heckscher-Ohlin country-industry measures of relative endowments of physical and human capital, bilateral trade costs variables from the CEPII GeoDist database (Mayer and Zignago, 2011), and fixed effects to account for exporter, importer and industry characteristics, as well as dummy variables that control for all omitted aggregated sector characteristics for the exporting and importing country. Robust standard errors are clustered at the exporter-importer level.

Sources: Bombardini, M., G. Gallipoli and G. Pupato (2012), “Skill dispersion and trade flows”, American Economic Review.

Bombardini, M., G. Gallipoli and G. Pupato (2009), “Skill dispersion and trade flows”, NBER Working Papers.

Mayer, T. and S. Zignago (2011), “Notes on CEPII’s distances measures: the GeoDist Database,” CEPII Working Paper 2011-25.

The main reason that has been put forward to explain why the dispersion of skills, and not only average skills, matter for trade is the degree of complementarity between the skills of a worker and the skills of any other team member with whom the worker is paired (Grossman, 2013). Industries differ in such complementarity of workers’ skills across tasks in the production process. Some industries, such as aerospace or engine manufacturing, require completing long sequences of tasks and poor performance at any single stage greatly reduces the value of output. These industries have high skills complementarity, known as the O-ring model (Kremer, 1993), in which efficiency improves when workers of similar skills are employed in every stage of production. In other industries, such as paper manufacturing, skills are more easily substitutable (low skills complementarity) and poor performance in some tasks can be mitigated by superior performance in others.

The degree of skills complementarity of an industry can be approximated by the degree of communication, contact and teamwork between workers in the industry (Bombardini, Gallipoli and Pupato, 2012). The more complementarity there is between a worker’s skills and the skills of other members of the team, the more they need to communicate with one another. In the Survey of Adult Skills, the questions covering these topics are summarised in the task-based skills indicator management and communication (see Box 3.1). According to this indicator, all complex business services and high-tech manufacturing industries show a high level of skills complementarity (Figure 3.9).

Industries requiring good performance at all stages of production because of the high complementarity of skills would benefit from a pool of reliable workers, or workers who perform at the expected level. In contrast, industries with low complementarity of skills can cope with workers with uneven skills. When recruiting, firms cannot fully observe the skills of applicants. However, they can observe a number of characteristics such as the level of education and training, and age, and they base their recruitment decisions on the basis of these observable skills determinants. Pools of reliable workers emerge in countries where individuals perform at the expected level or where individuals’ skills present no unwelcome surprises once their various characteristics have been accounted for, including education level. These countries have a small dispersion of unobservable skills. Overall, countries with smaller dispersions of unobservable skills have a trade comparative advantage in industries characterised by greater complementarities in the production process. Countries with a larger dispersion of unobservable skills have a trade comparative advantage in industries characterised by greater substitutability in the production process.

The Survey of Adult Skills shows how countries differ in their skills dispersion, for instance in terms of literacy skills (Table 3.2, first two columns).10 Several factors contribute to the population’s skills and therefore to the skills dispersion. Some are observable, such as the level of education, participation in training, age, and gender. However, individuals with similar observable characteristics do not have the same skills. In the same way, countries’ skills dispersions can be separated into two parts: one that comes from the dispersion of observable characteristics such as differences in education levels and the demographic structure; and one that cannot be explained by differences in observable characteristics, called the unobservable skills dispersion in Table 3.2. Countries do not have the same ranking in terms of the usual standard deviation of literacy scores and in terms of the unobservable skills dispersion. The unobservable skills dispersion can be large when there are differences in the quality of education programmes at the same level of education, or when characteristics that are more difficult to observe play an important role. Countries with a small unobservable skills dispersion can have pools of reliable workers in the sense that workers with the same observable characteristics would tend to perform at the same level.

Table 3.2. Characteristics of the literacy skills dispersion
2012 or 2015
Country	Standard deviation of Literacy scores		Unobservable skills dispersion
Country	Rank	Value	Rank	Value
Australia	24	50.47	22	0.18
Austria	5	43.96	6	0.15
Belgium	11	47.08	7	0.15
Canada	23	50.41	20	0.17
Chile	27	52.65	28	0.22
Czech Republic	3	40.79	4	0.14
Denmark	15	47.72	17	0.17
Estonia	7	44.40	9	0.15
Finland	26	50.67	8	0.15
France	20	49.02	19	0.17
Germany	14	47.40	10	0.15
Greece	9	46.65	25	0.18
Ireland	12	47.19	18	0.17
Israel	28	55.55	27	0.22
Italy	8	44.69	16	0.17
Japan	1	39.71	1	0.12
Korea	4	41.69	2	0.13
Netherlands	18	48.39	5	0.15
New Zealand	13	47.39	12	0.16
Norway	10	47.02	11	0.16
Poland	16	47.98	21	0.17
Slovak Republic	2	40.07	3	0.14
Slovenia	17	48.15	24	0.18
Spain	21	49.03	23	0.18
Sweden	25	50.56	14	0.16
Turkey	6	44.11	26	0.19
United Kingdom	19	48.97	15	0.17
United States	22	49.19	13	0.16
Note: All statistics are shown for the whole population in a country.
The unobservable skills dispersion is computed by: 1) estimating a regression of the logarithm of literacy scores on education, age, gender, immigration background and training; 2) computing the residuals of the regression for each individual (logarithm of literacy scores minus fitted values); 3) computing the standard deviation of the residuals by country.
Lower ranks indicate low skills dispersion and high mean while high ranks indicate high skills dispersion and low mean.
Chile, Greece, Israel, New Zealand, Slovenia and Turkey: Year of reference 2015. All other countries: Year of reference 2012. Data for Belgium refer only to Flanders and data for the United Kingdom refer to England and Northern Ireland jointly.
Source: OECD calculations based on the Survey of Adult Skills (PIAAC) (2012, 2015), www.oecd.org/skills/piaac/publicdata andanalysis.

Having pools of reliable workers (or a narrow unobservable skills dispersion) enables countries to have a comparative advantage in trade in industries characterised by a high skills complementarity (Figure 3.15). Countries have the highest differences in exports in value added terms with countries that show the highest difference in the unobservable skills dispersion. For instance, because Japan has a narrow unobservable skills dispersion, providing for pools of reliable workers, it can export (in value added terms) 23% more than Chile in industries with high skills complementarity (relative to industries with low skills complementarity). These results are symmetric: likewise, due to Chile’s large unobservable skills dispersion, its exports in industries with low skills complementarity (relative to industries with high skills complementarity) are 23% higher than those of Japan. As Korea also has a narrow unobservable skills dispersion, Japan’s exports in industries with high skills complementarity are only 2.6% higher than those of Korea.

Opportunities for specialisation

Countries vary in their skills endowments – the skills mix of their population and their unobservable skills dispersion, enabling them to have pools of reliable workers. At the same time, industries vary in their skills requirements. The interaction between countries’ skills endowment and industries’ specificities contributes to countries’ comparative advantages and enables countries to perform well in some GVCs.

Comparing countries’ potential for specialisation, arising from their skills endowment, with their actual industry specialisation and how it has evolved over the last decade can show how countries can specialise within GVCs by capitalising on their skills. This analysis assumes that all other potential sources of trade comparative advantage are held constant.

Countries’ specialisation within GVCs can be observed by looking at revealed comparative advantages (RCAs). The RCA indicates the relative advantage or disadvantage of a country in a certain class of goods or services, as evidenced by trade flows. The TiVA database makes it possible to compute RCAs in terms of value added and to capture countries’ specialisation in industries within GVCs. Traditionally, RCA analysis is based on comparing a country’s share of world exports of a particular product with its share of overall exports. However, the best way to determine specialisation within GVCs is to calculate RCAs on the basis of GVC incomes in the production of final goods, rather than exports, since the comparative advantage stems from primary factors of production in the value added and not the imported inputs. An RCA larger than 1 for an industry indicates that the share of the country’s overall GVC income that the country derives from adding value in the GVC production of this industry is higher than that of other countries.

Over the last 15 years, OECD countries have been increasingly specialising in services, while their RCAs in resource extracting sectors as well as in many manufacturing sectors have been declining (Table 3.3). Some variation exists, however. East European countries, along with Germany, Ireland, Israel and Korea, have enhanced their integration into high-tech industries, such as electrical and optical, or chemicals. Other countries, including Greece, Japan and the Netherlands, have increased RCA in low-technology sectors such as food products, and wood and paper.

Countries’ comparative advantages in GVCs resulting from their skills characteristics can be summarised by observing whether countries’ skills characteristics are aligned with industries’ skills requirements (Box 3.5). Different skills characteristics can provide comparative advantages in different industries. For instance, in terms of its skills mix, Israel could specialise in all high-tech manufacturing and complex business service industries, but its strong unobservable skills dispersion provides comparative advantages rather in low-tech and medium-tech manufacturing (Table 3.4).

Most OECD countries strive to reach technology frontiers and specialise in technologically sophisticated industries – either medium to high-tech manufacturing industries or in complex business services. Countries differ in the number of specialisation opportunities they may obtain from their skills characteristics (Table 3.5). Some countries (e.g. Estonia, Japan, Korea and New Zealand) could explore a wide spectrum of specialisation opportunities across the different technologically advanced sectors, while others have good skills alignment only in services (e.g. Austria, the Netherlands, Norway, the Slovak Republic and Slovenia) or manufacturing (e.g. Canada, Chile and Finland). Some countries’ skills characteristics struggle to meet the requirements of the technologically advanced sectors (Australia, Ireland, Turkey, the United Kingdom and the United States). Table 3.5. Specialisation opportunities in complex business services, high-tech and medium-high-tech manufacturing industries resulting from the alignment for countries’ skills characteristics with industries’ skills requirements.

Box 3.5. Deriving the specialisation opportunities stemming jointly from countries’ various skills characteristics

This chapter shows that several characteristics of skills can shape countries’ specialisation in GVCs: the skills mix, with a pure skills mix effect and a relative skills endowment effect, and the unobservable skills dispersion. This section consolidates the various results to assess the overall extent to which countries may have opportunities to specialise in complex business and high-tech manufacturing industries. These specialisation opportunities are then compared with the countries’ actual specialisations as measured by their RCAs.

According to the models discussed in this chapter, several skills characteristics shape specialisation opportunities: i) unobservable skills dispersion; ii) skills mix in numeracy versus literacy; iii) skills mix in problem solving in technology-rich environments versus literacy; and iv) skills mix in problem solving in technology-rich environments versus numeracy. The skills mix includes two aspects, a pure skills mix effect measured by the correlation between the relative skills advantage and the absolute skills advantage, and a relative skills endowment effect measured by the average ratio of each of two skills. The methodology includes several steps:

1) An overall skills mix effect combining the pure skill mix effect with the relative endowment effect is calculated based on the model specification for each of three pairs of skills (Grundke et al., forthcoming b).

2) An average unobservable skills dispersion for literacy, numeracy and problem solving in technology-rich environments is calculated.

3) A measure of the alignment of countries’ skills characteristics with industry’ skills requirements is calculated for each country-industry pair, and for each skill characteristic. This is done by ranking countries in terms of the four skills characteristics and industries in terms of their skills requirements and looking at the distance between these ranks. The smaller the distance, the more aligned a country’s skills are with this industry’ skills requirements. For instance, the country with the lowest unobservable skills dispersion has the lowest distance (strongest alignment) with the industry with the highest intensity in managing and communication tasks, which indicates a strong degree of skills complementarity. The country with the strongest skills mix in numeracy versus literacy (both a high correlation between the relative numeracy skills and the absolute literacy, and high relative numeracy skills) has the lowest distance (strongest alignment) with the industry with the highest relative intensity in marketing and accounting tasks versus managing and communication ones.

4) An overall alignment measure of countries’ skills characteristics with the skills requirements of a particular industry is calculated by taking the average of the alignments in terms of the four skills characteristics and industry’s skills requirements. This number would measure the joint alignment.

5) A threshold needs to be applied to decide about the minimum degree of alignment that leads to an opportunity to specialise in the industry. Only countries with an alignment that is in the top 25 percentile of the alignment distribution across all countries and industries in the sample are considered to have a specialisation opportunity.

Table 3.5 shows the specialisation opportunities stemming from countries’ skills characteristics and countries’ current specialisations as reflected by their RCAs.

Source: Grundke, R. et al. (forthcoming b), “Having the right mix: The role of skills bundles for comparative advantage and industry performance in GVCs”, OECD Science, Technology and Industry Working Papers.

Table 3.4. Countries’ comparative advantages in global value chains in various types of industries stemming from their skills characteristics
Skills requirements of industries	Complementarity (1)		Intensity of marketing/accounting versus managing/communication skills (2)				Intensity of self-organisation skills versus managing/communication skills (3)				Intensity of self-organisation skills versus marketing/accounting skills (4)
	High	Low	High		Low		High		Low		High		Low
Examples of industries	All complex business services and high-tech manufacturing	Most low-tech and medium-tech manufacturing and less complex business services	Most business services and low-tech manufacturing		Most high-tech and medium-tech manufacturing industries		Some complex business services and various manufacturing		All less complex business services and various manufacturing		High-tech and medium-tech manufacturing		Most business services and low-tech manufacturing
Channel	Unobservable dispersion	Unobservable dispersion	Mix	Endowment	Mix	Endowment	Mix	Endowment	Mix	Endowment	Mix	Endowment	Mix	Endowment
Australia		**	**			**			**	*		**	**
Austria	**			**	*			**	**				*	*
Belgium	**			**	**		**			**	**			**
Canada		*	*			*	*	*				*	*
Chile		**			*	**		**	**			**	**
Czech Republic	**			**	**		**	*			**			*
Denmark		*		**	**		*	**			*			**
Estonia	*			*	**		*			**	**			**
Finland	*			*	**				*	**	*			*
France		*	**			*
Germany	*		*	**				**	*			*	*
Greece		**		*	**			*	*		*			*
Ireland		*	**			**			**	*		*	**
Israel		**	*			*	*	**			*	*
Italy		*		*	*
Japan	**				*	*	**			**	**			**
Korea	**		*			*	**			*	*	*
Netherlands	**			*	*				*	**	*	*
Norway	*		**	*				*	**			*	**
New Zealand	*		*			**	*	*				**	*
Poland		*			**	*	**			**	**			*
Slovak Republic	**		*	**			*			*			*	**
Slovenia		**	*	**			**	*			**			**
Spain		**			*	*
Sweden	*			*	*				*	*	*			*
Turkey		**	**			**		**	*			**	*
United Kingdom		*	**			**		*	**			**	**
United States	*		**			**	*			*		**	**
Note: Results come from the models presented in Boxes 3.3 and 3.4. ** and * indicate that countries are among the 75th percentile or 50th percentile of the exporters in value added terms in the selected type of industries given the characteristics of their skills distribution.
(1) GVC comparative advantages are given by countries’ unobservable skills dispersions in literacy.
(2) GVC comparative advantages are given by countries’ correlations between comparative advantage of individuals in numeracy and absolute advantage in literacy for the skills mix channel and by the average ratio of scores in numeracy and literacy for the skills endowment channel.
(3) GVC comparative advantages are given by countries’ correlations between comparative advantage of workers in problem solving in technology-rich environments and absolute advantage in literacy for the skills mix channel and by the average ratio of scores in problem solving in technology-rich environments and literacy for the skills endowment channel.
(4) GVC comparative advantages are given by countries’ correlations between comparative advantage of workers in problem solving in technology-rich environments and absolute advantage in numeracy for the skills mix channel and by the average ratio of scores in problem solving in technology-rich environments and numeracy for the skills endowment channel.
Chile, Greece, Israel, New Zealand, Slovenia and Turkey: Year of reference 2015. All other countries: Year of reference 2012. Data on skills for Belgium refer only to Flanders and for the United Kingdom – England and Northern Ireland jointly.
Source: OECD calculations based on the Survey of Adult Skills (PIAAC) (2012, 2015), www.oecd.org/skills/piaac/publicdataandanalysis; OECD Trade in Value Added database (TiVA), https://stats.oecd.org/index.aspx?queryid=66237; OECD Annual National Accounts, SNA93, http://stats.oecd.org/; OECD STAN STructural ANalysis Database, http://stats.oecd.org/; Mayer and Zignago (2011), “Notes on CEPII’s distances measures: the GeoDist Database”, CEPII Working Paper 2011-25; World Input-Output Database (WIOD), www.wiod.org/home.

Comparing countries’ current specialisations in GVCs, as captured by the RCAs, with the specialisation opportunities emerging from countries’ skills characteristics leads to a number of findings:

Several countries have increased their RCAs in industries in which their skills characteristics give them specialisation opportunities (e.g. Japan in the computer industry). This indicates that their skills policies are in line with their specialisation patterns and objectives in GVCs.
Some countries have increased their RCAs in some industries but this change is not supported by their skills characteristics (e.g. Canada and the United States in most complex business services). The poor alignment of their skills with industry requirements could make it difficult for them to maintain their comparative advantage.
Other countries have specialisation opportunities in some industries because of their skills characteristics but have seen their RCAs decrease in these industries (e.g. Sweden in electrical machinery). This could be because it is no longer optimal to specialise in these industries; the countries concerned might be upgrading to other industries. Another explanation could be that factors other than skills prevent these countries from specialising in these industries.
Finally, in some cases skills characteristics do not provide specialisation opportunities and countries’ RCAs have decreased (e.g. Australia, Canada, Norway and the United Kingdom, in several high-tech industries). This could indicate that countries are now specialising in other industries, such as services. But it could also indicate that to specialise in these industries, countries need to upgrade the skills of their population and achieve a better alignment of their populations’ skills with the skills requirements of these industries.

A closer look at how well countries’ different skills characteristics match the skills requirements of the technologically most advanced industries can indicate what countries could do in terms of skills to achieve specialisation objectives (Figure 3.16). Some patterns emerge:

A first group of countries have a strong alignment of their skills mixes with the requirements of these industries, but a large unobservable skills dispersion prevents them from having the pools of reliable workers that are required in these industries (Canada, Chile, Greece, Israel, Poland, Slovenia and Turkey). Countries in this group need to narrow their unobservable skills dispersion and improve or maintain a good skills mix to increase or strengthen their specialisation in technologically advanced industries. Israel has the strongest alignment in terms of its skills mix, but the lowest in terms of the unobservable skills dispersion. Countries in this group have different skills mixes, with Israel, Slovenia, Poland and New Zealand having strong skills mixes in problem solving in technology-rich environments: those who have strong problem solving skills relative to other skills also have strong numeracy and literacy skills. A strong skills mix in problem solving in technology-rich environments is required by several high-tech manufacturing industries and complex business services.
A second group of countries (Australia, Ireland, the United Kingdom and the United States) has a poor alignment of their skills characteristics – mainly the skills mix but also to some extent the unobservable skills dispersion – with the requirements of technologically advanced industries. These countries would need to develop stronger skills mixes and narrow the unobservable skills dispersion to increase or maintain comparative advantages in these industries.
The largest group of countries is characterised by a small unobservable skills dispersion, enabling them to have pools of reliable workers, and skills mixes that broadly correspond to the requirements of technologically advanced industries. This good overall alignment of skills characteristics with skills requirements of technologically advanced industries brings them some opportunities for specialisation in one or several of these industries. However, there are differences among countries. New Zealand would have to narrow its unobservable skills dispersion to increase or strengthen its specialisation in these industries and Austria, Denmark and Norway would have to develop stronger skills mixes.

Policies can help countries to develop strong skills mixes that meet the requirements of technology advanced industries. They can also narrow the unobservable skills dispersion. Large unobservable skills dispersion can reflect various factors. If the quality of education programmes varies widely at the same level of education, individuals can have the same type of formal diploma but different levels of skills. A segmented economy, in which leading firms offer a lot of non-formal training and access to the latest technologies while other firms are lagging behind in terms of knowledge, would also create differences in workers’ skills. As a result, individuals with a similar profile do not perform at the same level, creating unwelcome surprises for employers and weakening the efficiency of the production process. Policies can reduce the unobservable skills dispersion, either ex-ante, for instance through an education system with homogenous quality, or ex-post, for instance through measures to better signal individuals’ skills. Training policies can help those who do not perform at the expected level to catch up. Chapter 4 discusses these policies.

The results presented above come from models that have been estimated, that rely on a number of assumptions and that are constrained by data availability. They use information from the Survey of Adult Skills to assess the impact of specific skills – literacy, numeracy and problem solving in technology-rich environments – on countries’ opportunities to specialise in some industries. However, the results that show the importance of having a strong skills mix and a small dispersion of unobservable skills go beyond the set of skills assessed in the Survey of Adult Skills. A strong skills mix also means having strong cognitive skills and personality traits.

Due to data limitations, this analysis includes a small group of exporting countries, those covered by the Survey of Adult Skills. It shows the extent to which a country from this group can export more in value added terms (to the world) than another country of the same group because of its skills characteristics. The analysis does not enable the comparative advantage stemming from the skills characteristics of one country of this group to be compared with the advantage of a country outside the group, such as China. However, the results on the revealed comparative advantages and their evolutions do include all countries in the world.

Summary

A broad range of skills matter for participation and specialisation in GVCs. They include cognitive skills, personality traits and skills that combine both, such as the capacity to interact and communicate with others. Countries with the highest skills levels also participate and export the most in GVCs.

Countries can shape their specialisation in GVCs by developing skills characteristics that match industries’ skills requirements. The results here do not specify which skills characteristics countries should develop to make the most of GVCs. However, the results do illustrate the potential costs of adopting industry specialisation objectives that are misaligned with countries’ skills. Policies to support a specific industry can be inefficient if countries’ skills do not match the skills requirements of the industry and, by leading to misallocations of skills, they can lower the comparative advantage countries have in other industries.

Many OECD countries strive to excel in technologically advanced sectors, but the specialisation pathways for some countries would require more effort and take longer depending on their current production structure, skills characteristics and other capabilities. The more capabilities two industries share, the more likely it is that a country that successfully creates value in one of these industries will also specialise in the other. Some countries that lack the skills characteristics necessary for high-tech industries and complex business services can specialise first in industries that use available skills, while developing the necessary mix of skills. Other countries have skills characteristics that bring them opportunities to specialise in sophisticated industries. However, other factors might prevent them from specialising in these industries. For specialisation strategies to succeed, skills policies need to be implemented in line with other types of policies.

Industries vary greatly in their skills requirements. However, even industries with a low level of technological sophistication require a broad range of skills. The empirical analysis shows the importance of having the right mix of skills to perform in GVCs. These mixes of skills are specific to industries but they all involve various cognitive skills and “soft skills”. Education and training policies, for youth and adults, students, workers and the unemployed, are crucial to develop these skills mixes. The population needs to have skills that match industries’ skills requirements, as well as other types of skills that are required at the international level. This finding has implications for the design of educational programmes, especially for those that aim to develop advanced skills in a particular area, such as STEM programmes.

Countries can specialise in technologically advanced industries if they are able to provide pools of reliable workers who perform at the expected level, which requires a narrow dispersion of unobservable skills. Some OECD countries show a large unobservable skills dispersion, however. Ex-ante policies to achieve equal quality across similar educational programmes and ex-post policies to train workers who do not perform at the expected level, or to better signal workers’ skills, are crucial to enable countries to specialise in complex services and high-tech manufacturing industries.

References

Asuyama, Y. (2012), “Skill distribution and comparative advantage: A comparison of China and India”, World Development, Vol. 40/5, pp. 956-969.

Barney, J. (1991), “Firm resources and sustained competitive advantage”, Journal of Management, Vol. 17/1, pp. 99-120.

Bharadwaj, A.S. (2000), “A resource-based perspective on information technology capability and firm performance: An empirical investigation”, MIS Quarterly, Vol. 24/1, Management Information Systems Research Center, University of Minnesota, pp. 169-96, https://doi.org/10.2307/3250983.

Bloom, N. and J.V. Reenen (2010). “Why do management practices differ across firms and countries?”, The Journal of Economic Perspectives, Vol. 24/1, pp. 203-224.

Bloom, N. et al. (2012), “Management practices across firms and countries”, The Academy of Management Perspectives, Vol. 26/1, pp. 12-33.

Bombardini, M., G. Gallipoli and G. Pupato, (2012), “Skill dispersion and trade flows”, American Economic Review, Vol. 102/5, pp. 2327-2348.

Bombardini, M., G. Gallipoli and G. Pupato (2009), “Skills dispersion and trade flows”, NBER Working Papers, No. 15097, The National Bureau of Economic Research, Cambridge, MA.

Bougheas, S. and R. Riezman (2007), “Trade and the distribution of human capital”, Journal of International Economics, Vol. 73/2, pp. 421-433.

Chor, D. (2010), “Unpacking sources of comparative advantage: A quantitative approach”, Journal of International Economics, Vol. 82, pp. 152-167.

Conti, G. et al. (2014), “Bayesian exploratory factor analysis”, Journal of Econometrics, Vol. 183/1, pp. 31-57.

Costello, A.B. and J.W. Osborne (2005), “Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis”, Practical Assessment, Research & Evaluation, Vol. 10.

Goldberg, L.R. (1990), “An alternative ’description of personality’: The big-five factor structure”, Journal of Personality and Social Psychology, Vol. 59/6, pp. 1216-1229.

Grossman, G. (2013), “Heterogenous workers and international trade”, NBER Working Paper, No. 18788, The National Bureau of Economic Research, Cambridge, MA.

Grossman, G. (2004), “The distribution of talent and the pattern and consequences of international trade”, Journal of Political Economy, Vol. 112, pp. 209-39.

Grossman, G. and G. Maggi (2000), “Diversity and trade”, American Economic Review, Vol. 90/5, pp. 1255-1275.

Grundke, R. et al. (forthcoming a), “Skills and global value chains: Characterisation and evidence”, OECD Science, Technology and Industry Working Papers, OECD Publishing, Paris.

Grundke, R. et al. (forthcoming b) (2017b), “Having the right mix: The role of skills bundles for comparative advantage and industry performance in GVCs”, OECD Science, Technology and Industry Working Papers, OECD Publishing, Paris.

Ichniowski, C., K. Shaw and G. Prennushi (1997), “The effects of human resource management practices on productivity: A study of steel finishing lines”, American Economic Review, Vol. 87, pp. 291-313.

Hanushek, E.A. and L. Woessmann (2008), “The role of cognitive skills in economic development”, Journal of Economic Literature, Vol. 46/3, pp. 607-668.

Heckman, J.J. and T. Kautz (2013), “Fostering and measuring skills: Interventions that improve character and cognition”, NBER Working Paper, No. 19656, The National Bureau of Economic Research, Cambridge, MA.

Heckman, J.J. and Y. Rubinstein (2001), “The importance of noncognitive skills: Lessons from the GED testing program”, American Economic Review, Vol. 91/2, pp. 145-149.

Heckman, J.J., J. Stixrud and S. Urzua (2006), “The effects of cognitive and noncognitive abilities on labor market outcomes and social behaviour”, Journal of Labor Economics, Vol. 24/3, pp. 411-482.

Hitt, M., R.D. Ireland and R. Hoskisson (2012), Strategic Management Cases: Competitiveness and Globalization, Cengage Learning, Boston, MA.

Hoyles, C. et al. (2002), “Mathematical skills in the workplace: Final report to the Science Technology and Mathematics Council”, Institute of Education, University of London.

Humburg, M., R. van der Velden and A. Verhagen (2013), “The employability of higher education graduates: The employers’ perspective”, Publications Office of the European Union, Luxembourg.

Kautz, T. et al. (2014), “Fostering and measuring skills: Improving cognitive and non-cognitive skills to promote lifetime success”, NBER Working Paper, No. 20749, The National Bureau of Economic Research, Cambridge, MA.

Kremer, M. (1993), “The O-ring theory of economic development”, Quarterly Journal of Economics, Vol. 108/3, pp. 551-75.

Levchenko, A.A. (2007), “Institutional quality and international trade”, Review of Economic Studies, Vol. 74/3, pp. 791-819.

Mayer, T. and S. Zignago (2011), “Notes on CEPII’s distances measures: the GeoDist database”, CEPII Working Paper, No. 2011-25.

Mohan, G. and Z.R. Mulla (2013), “Openness to experience and work outcomes: exploring the moderating effects of conscientiousness and job complexity”, Great Lakes Herald, Vol. 7/2, pp. 18-36.

Nunn, N. (2007), “Relationship-specificity, incomplete contracts, and the pattern of trade”, The Quarterly Journal of Economics, Vol. 122/2, pp. 569-600.

Ohnsorge, F. and D. Trefler (2007), “Sorting it out: International trade with heterogeneous workers”, Journal of Political Economy, Vol. 115/5, pp. 868-892.

Romalis, J. (2004), “Factor proportions and the structure of commodity trade”, American Economic Review, Vol. 94/1, pp. 67-97.

Santhanam, R. and E. Hartono (2003), “Issues in linking Information Technology capability to firm performance”, MIS Quarterly, Vol. 27/1, Management Information Systems Research Center, University of Minnesota, pp. 125-53, www.jstor.org/stable/30036521.

Schmidt, F.L. (2002), “The role of general cognitive ability and job performance, why there cannot be a debate”, Human Performance, Vol. 15/1-2, pp. 187-210.

Schmidt, F.L. and J. Hunter (2004), “General mental ability in the world of work: Occupational attainment and job performance”, Journal of Personality and Social Psychology, Vol. 86/1, pp. 162-173.

Tippins, M.J. and S.S. Ravipreet (2003), “IT competency and firm performance: is organizational learning a missing link?”, Strategic Management Journal, Vol. 24/8, pp. 745-761.

UKCES (2011), “The supply of and demand for high level STEM skills”, Briefing Paper, December 2011, UK Commission for Employment and Skills (UKCES).

Wright, P., G. McMahan and A. McWilliams (1994), “Human resources as a source of sustained competitive advantage”, International Journal of Human Resource Management, Vol. 5, pp. 299- 324.

Notes

← 1. The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

← 2. Except the readiness to learn indicator.

← 3. Except for the 25th percentile with backward linkages in terms of exports.

← 4. For instance, business services industries are intensive in ICT and STEM skills but have on average low backward participation. Likewise, the use of ICT and STEM skills is increasing with the occupation category, with managers and professionals using these skills the most while their jobs may also be less exposed to offshoring than other jobs.

← 5. These results are in contrast with the literature, according to which these skills are crucial for performance in a complex global environment. Indeed, globalisation calls for an ever greater ability to adapt to change and absorb shocks. Hence, self-organisational capability and workers’ flexibility should correspondingly lead to superior firm performance in GVCs. Marketing skills can also increase firms’ ability to participate in GVCs as they are needed to look outside their existing business context, to develop new perspectives on managing products, work with new distributors and suppliers and reach new customers and competitors.

← 6. Workers may have the skills but not use them in their jobs. This argument could explain why all skills indicators based on the information on the frequency of tasks performed (the so-called task-based skills) show weaker links with GVC variables, as compared with the assessed cognitive skills.

← 7. Since the intensity in ICT and STEM tasks does not appear to characterise groups of industries, the relative intensities with respect to these two tasks are not shown in Figure 13.

← 8. On the basis of an article published in The New York Times on 17 October 2006, www.nytimes.com/2006/10/17/world/asia/17india.html.

← 9. In Chile and the Turkey, a large share of adults opted out of the computer-based assessment or failed the ICT core or had no computer experience, which might partly explain why those who did the assessment have better problem solving skills in technology-rich environments than literacy or numeracy skills.

← 10. Literacy is defined in the Survey of Adult Skills as “understanding, evaluating, using and engaging with written text to participate in society, to achieve one’s goals, and to develop one’s knowledge and potential”. As such, compared with the other two assessed skills (numeracy and problem solving in technology-rich environments), literacy can be considered as the universal skill, the major prerequisite to find and maintain a job and contribute to a firm’s and a country’s performance.