Skill content of intra-European trade flows.

• Author: Zeddies, Goetz
• Position: Report

1. Introduction

   In the course of globalization, highly-developed countries should have increasingly specialized in human capital-intensive manufactured goods and, in return, should have sourced labor-intensive manufactured products from low-wage countries. For this reason, especially the opening up of Eastern Europe, but also the international integration of the Newly Industrializing East Asian Economies is often considered a significant cause of labor demand shifts detrimental for the lower-qualified in Western European countries, since international trade should favor the high-skilled in these countries (e.g. Freeman 1995, Wood 1995). This paper addresses this question by analyzing the skill content of bilateral intra-European trade flows of selected EU Member States, allowing inferences to be made about the impact of these trade relations on factor demand patterns in those countries. Previous studies on the factor content of trade have shown that, even in trade between highly developed countries, the results depend largely on whether or not international differences in technology are considered. Unlike previous studies, this paper takes up this issue by analyzing the high-, medium- and low-skill content of bilateral trade between Western European and, for comparison, also between Western and Eastern European countries. The skill content of trade is analyzed for identical, but also for different technologies by using country specific input-output and factor input data. According to theory, different factor intensities in production are a prerequisite for vertical product differentiation, which has become more and more important in recent decades (Falvey and Kierzkowski 1987, Helpman 1981). As the analyses show, the results do indeed depend largely on whether vertical product differentiation is allowed for or not.

   The paper is structured as follows: The following section contains a review of the literature, while section 3 describes the theoretical background and the model used in calculating the skill content of trade. In section 4, the skill content of the bilateral trade flows of six Western European countries is calculated by applying input-output techniques. Finally, section 5 closes with some concluding remarks.

2. Review of the Literature

   One of the main theoretical foundations for explaining international trade patterns and their consequences for factor demand and income distribution in trading partner countries is the neo-classical Heckscher-Ohlin (HO) model of trade. According to this model, each country will specialize in and export commodities utilizing its abundant and thus comparatively cheap factors of production and will import goods using its scarce factors of production. An empirical test of the HO-theorem for the United States performed by Leontief (1953) seemed to disprove the hypothesis that countries' patterns of specialization are determined by factor proportions. In a model with two production factors (capital and labor), Leontief disaggregated the US economy into 50 industries, 38 of which produced tradable goods. He showed that in 1947, US imports were 30% more capital-intensive than US exports, although at the time the US was considered to be one of the most capital-abundant countries in the world. Today it is widely accepted that, besides trade barriers, differences in labor force qualifications is the main reason for this Leontief paradox (Baldwin 1971, Kravis 1956, Trefler 1993).

   So far, analyses investigating the factor content of trade for different countries exist (e.g. Dasgupta et al. (2009) for India, Engelbrecht (1996) for Germany, Webster (1993) for the UK and Widell (2005) for Sweden), as well as studies testing traditional trade theories (e.g. Bowen et al. 1987, Davis and Weinstein 2001, Maskus 1985, Staiger 1988, Trefler 1995). However, in many of these last mentioned studies the empirical results of these tests are quite inconclusive. A critical concern is that the bulk of these studies assume identical production technologies and factor inputs across countries for calculating the factor content of countries' trade. As a consequence, the factor content of exports and imports hardly deviates one from the other. However, according to New Trade Theories, identical production technologies would imply only horizontal product differentiation, resulting in horizontal intra-industry trade. In this case, imports would differ from domestically manufactured export goods only with respect to product characteristics, but would be of the same quality. But these days, product differentiation is largely vertical, which means that goods are manufactured with different factor proportions or technologies and differ with respect to quality and prices (Falvey and Kierzkowski 1987, Flam and Helpman 1987). By implication, this means that producing a perfect import substitute would require exactly the same factor inputs and production technology that are applied when producing the considered product abroad. If not, the domestically manufactured import substitute and the imported product would not be homogeneous. Against this background, quantifying domestic job losses induced by imports for different skill groups requires calculating the factor content of imports by using technology as well as factor input matrices of trading partner countries.

   Although empirical analyses investigating the factor content of trade on a bilateral level have more recently used technology matrices of both the exporting and the importing country (e.g. Choi and Krishna 2004, Davis and Weinstein 2003, Harrigan 1997, Lundberg and Wiker 1997, Nishioka 2006 and Torstensson 1992), many of these analyses are restricted to highly developed OECD countries, which probably share quite similar production technologies and factor endowments. Although Hakura (1999) found that theoretical hypotheses are empirically supported for EU Member States if different technology matrices are used for the countries considered, only bilateral trade relations between the high income Western European countries of Belgium, Germany, France, Italy and the Netherlands were considered. For EU member states, only Cabral et al. (2006 and 2009) focused on trade between high-income countries (the UK and others, respectively) and middle-income countries. However, for the former, only the United Kingdom's, and for the latter, only the Portuguese technology matrix was used and considered as representative.

   Against this background, in this study trade flows between not only selected Western, but also between Western and new Eastern European EU member states will be taken into account. The countries considered are Austria, Denmark, France, Germany, the Netherlands and Sweden and, as Eastern European trading partners, the Czech Republic, Hungary, Poland and Slovakia. Thereby, those Western European countries with the most intensive trade relations with Eastern Europe, measured by the share of the four Eastern European countries in total exports and imports, are considered. With respect to Eastern European countries, the country selection was limited by data availability. Detailed data on labor input by industry are only available for the four countries mentioned above. However, these four countries are the most important Eastern European trading partners of the Western European countries considered in this study. Since data on capital input are only available for the Czech Republic and Hungary, capital was not taken into account. (2) Unlike many other studies dealing with factor content in bilateral trade (e.g. Davis and Weinstein 2001, Harrigan 1997, Lai and Zhu 2007), in this study the total labor force will be subdivided into human capital and lower-qualified labor in order to deduce factor demand patterns arising from international trade between EU Member States. This will be done by identifying the high-, medium- and low-skill content of intra-European trade flows. In this way, the calculations will be performed in the case of identical as well as of different technologies across countries by using national factor input and input-output matrices. This allows for a consideration of country specific factor inputs resulting from endowment differences.

3. Theoretical Background and Specification of the Model

   3.1 The HOV-Model of Trade

   According to the Heckscher-Ohlin model of trade, international trade results from differences in countries' factor endowments. A modification of the traditional HO model suggests that, under the assumption of balanced trade, identical production technologies, identical and homothetic preferences across countries, no factor intensity reversals and free trade, international trade will accomplish the task of exchanging the services of production factors embodied in tradable goods and services (Vanek 1968). This Heckscher-Ohlin-Vanek (HOV) version of the classical HO model therefore implies that countries should have a net export of relatively abundant factor services and a net import of relatively scarce factor services (Melvin 1968). In the following paragraphs, the HOV model will be formally derived. (3)

   Besides direct factor inputs, the production in industry i normally requires intermediate inputs from other industries in order to produce country m's gross output ([Y.sup.m.sub.gross]). These are captured by the (i x i) input-output matrix of country m, which can be easily transformed into a technical coefficients matrix, denoted by [A.sub.m]. Each element in [A.sub.m] shows the units of input from different industries necessary for producing one unit of output in industry i. Under the presence of intermediate inputs, the interrelationship between gross and net output ([Y.sup.m.sub.net]) of country m is expressed by,

   [Y.sup.m.sub.net] = (I - [A.sup.m])[Y.sup.m.sub.gross] (1)

   where I represents the (i x i) identity matrix. Assuming that (I - [A.sub.m]) is invertible, a (f x i) matrix of total (direct and indirect)...