ETSG 2006 Vienna Universität Wien, 7-9 September. Vertical and Horizontal Intra-Industry Trade in Japan: Patterns and Their Determinants

Transcription

1 Preliminary draft ETSG 2006 Vienna Universität Wien, 7-9 September Vertical and Horizontal Intra-Industry Trade in Japan: Patterns and Their Determinants Takamune FUJII * Faculty of Business, Aichi University First version: 01/09/2006 Abstract: According to the theoretical literature, causes of intra-industry trade differ by types, vertical and horizontal IIT. If the theoretical implications are plausible, differences in factor intensity or fragmentation in production should yield vertical IIT, and product differentiation or monopolistic competition should yield horizontal IIT. In this paper, we analyze the pattern of vertical / horizontal intra-industry trade in Japan by using CEPII index by trade partner and industry. And then, we investigate the determinants of both types of IIT and examine the relevancy of theoretical implication. Keywords: Vertical/Horizontal Intra-industry trade, Monopolistic Competition, Quality ladder, Love of variety, Fragmentation JEL Classification: F11, F12, F14 * 370 Kurozasa, Miyoshi-cho, Nishikamo-gun, Aichi Japan. TEL: , FAX: , tkfujii@vega.aichi-u.ac.jp

2 Vertical and Horizontal Intra-Industry Trade in Japan: Patterns and Their Determinants Takamune FUJII ** Faculty of Business, Aichi University 1. Introduction In recent years, the academic literature of intra-industry trade enjoys prosperity again. Various theoretical models are developed to explain the patterns and reasons of IIT, and empirical methods are improved to inspect the effectiveness of theoretical models. Especially, mainly European academic scene, many researchers try to classify the IIT as two types, vertical IIT and horizontal IIT, to understand the patterns of IIT more accurately and specify the exact determinants of IIT, and then many interesting results are found by their studies. However, there are relatively scarce studies to classify the types of IIT in Japan, then the exact patterns and determinants of IIT in Japan are ambiguous. In this paper, based on the these methodological improvement, we try to classify Japanese IIT in two types, vertical IIT and horizontal IIT, and examine the patterns and determinants of these two types of IIT by using simple econometric model. The following sections are as follows. In the next section, we summarize and classify the many theoretical researches for explaining IIT, and examine the measurement techniques of IIT in section 3. Section 3 also shows brief summary of the patterns of Japanese IIT. In section 4, simple empirical method is developed and some hypotheses about the determinants of two types of IIT are considered. The results of empirical analysis are reviewed in section 5, and concluded in section Theoretical perspectives of VIIT/HIIT The theoretical literature for explaining the IIT is classified roughly two types. First pattern is vertical intra-industry trade. In this literature, intra-industry trade is occurred among different quality goods or different producing process, like as the parts/components and final goods. Because comparative advantage structure is influencing this type of IIT, the determination of the trade can be illustrated by expanding traditional trade theory. For example, Finger (1975), Lipsey (1976), Raymond (1976) and Falvey (1981) developed traditional trade model with many goods and ranked by their quality. In their models, intra-industry trade among different quality goods are explained by the difference of factor intensity between two countries. These models are so called quality ladder model. Another pattern of the vertical IIT is so called fragmentation. In this pattern, ** The author thanks to participants of the NIESG (Nagoya International Economics Study Group) workshop, held in last year in Chukyo University, for helpful comments.

3 one country exports the component of the goods and counterpart exports final goods. This case of IIT is also fully explained by the patterns of factor intensity or technological level of two countries 1. Second patter of IIT is so called horizontal IIT, and cannot be explained by the traditional comparative advantage structure. The keys of this type are monopolistic competition and product differentiation. There are two types of theoretical explanations, ideal variety model and love of variety model. Both models based on the monopolistic competition, former model hypothesizes that the consumers want to buy the good which has the most ideal characteristics for them. 2 Latter model hypothesized that consumers want to buy many variations of the goods and gain welfare from the amount of variety 3. To summarize the findings of the studies, there are two types of theories for explaining IIT and the determinants of IIT is truly different between two types of theory 4. In other words, we may have to divide the IIT index as two types for accurate analysis. In the next section, we examine the measurement technique IIT and try to divide it as two types. 3. Measuring the VIIT/HIIT and the overview of IIT in Japan 3.1. Measurement methods According to the theoretical literature, we should strictly distinguish vertical IIT and horizontal IIT when we measure the IIT index empirically. In recent empirical studies, some measurement techniques are established to distinguish vertical and horizontal IIT, and many researches are conducted in various countries. There are roughly two ways to measure the vertical/horizontal IIT, one is the extension of Grubel-Lloyd Index 5, the other is trade overlapping index or CEPII index. Former is a method that is adopted by Greenaway, Hine and Milner (1994, 1995) at first, and the latter is that classify the amount of trade itself by using the differences of unit price between exported commodities and imported commodities, adopted firstly by Fontagne and Freudenberg (1997) and Fontagne, Freudenberg and Peridy (1997). Mainly because of the convenience for calculations, we adopt latter method (hereafter CEPII index) in this paper for distinguishing the Japanese vertical/horizontal IIT. The CEPII index is calculated the degree of import and export overlap in each commodity, and distinguished trade amount of total sector by using unit price of export and import commodity 6. 1 For more details, please refer Arndt and Kierzkowski (2001) and Cheng and Kierzkowski (2001). 2 For instance, Lancaster (1980) and Helpman (1981) 3 For instance, Krugman (1979, 1980) and Dixit and Norman (1980) 4 There are some other types of explanation theory for IIT, for instance, reciprocal dumping model (Brander (1981), Brander and Krugman (1983)), and so on. However, in my opinion, these models have little reality, and it is no mean to examine reciprocal dumping model empirically. 5 For the calculation strategy of typical Grubel-Lloyd index, see Grubel and Lloyd (1975) and so on. 6 Abd-el-Rahman (1991) is one of the first researches to use unit price as an index of the quality differential of traded goods. Stiglitz (1987) justified theoretically that the unit price difference reflects the quality difference between goods/commodities.

4 At first, we define the degree of the trade overlap is an traded goods as follows, TOL i = Min(X i,m i ) Max(X i,m i ). (1) TOL i is the degree of trade overlap, X i is total value of export, and M i is total value of import in sector i, respectively. Then, if TOL i is equal or less than 10%, we regards the trade pattern in this sector as one-way trade, and TOL i is more than 10%, we regards as two-way trade, in other words, intra-industry trade. In case of one-way trade, the comparative advantage structure is the main engine of the trade, told as traditional trade theory. However, in case of two-way trade, there may be another engine of trade. Next, we measure the unit price of all exported commodities and imported commodities in the sector, and classify the trade amount of two-way trade in the sector. If the unit value of a commodity j in sector i is as follows, the trade pattern of this commodity is defined as horizontal IIT. x 1 1+ UV ij UV ij m 1+ (2) And, if the unit value is as follows, the trade pattern is defined as vertical IIT. x 1 1+ > UV ij UV or 1+ < UV ij m ij UV (3) m ij UV x ij, UV m ij is the unit value of exported commodity j in sector i, and unit value of imported commodity, respectively, and denotes some constant 7. At last, the share of the amount of horizontal IIT and vertical IIT in the sector is calculated and used as the indices of horizontal/vertical IIT. These two methods, expanded Grubel-Lloyd index and CEPII index, are adopted by various studies for measuring and analyzing the recent IIT patterns in various countries Brief overview of Japanese IIT In this section, we review briefly the trade pattern between Japan and rest of the world in Following the Fontagne et al (1997), we classify total trade into three types, one-way trade, vertical IIT and horizontal IIT in HS 6-digit commodity level and calculate the share of each type of trade in 2-digit sector level. The data is derived from the PC-TAS (UNCTAD). Table 1 shows main one-way trade sectors in Japan. Most of the sectors are primary products and processed goods. x 7 This constant mainly represents the range of homogeneity, or, more strictly, the differential between fob export price and cif import price. The value of this constant is various in many studies, but most studies use 0.15 to 0.3 as this constant. 8 For the example of the studies using expanded Grubel-Lloyd index, there are Aturupane, Djankov and Hoekman (1999) for EU and CEEC trade, Hu and Ma (1999) for China, Greenaway, Milner and Elliott (1999) for UK and EU, Blames and Martin (2000), Martin-Montaner and Rios (2002), Crespo and Fontoura (2004) for Spain and Portugal, Veeramani (2002) for India, and Mora (2002) for more classification the vertical IIT. For the example of the studies used CEPII index, there are Ishida (2002), Fukao, Ishido and Ito (2003), Kimura and Ando (2005) and Ando (2006) mainly for Japan.

5 < insert table 1 > Tables 2 and 3 show the sectors that the share of IIT is high. Table 2 presents vertical IIT, and table 3 presents horizontal IIT. The constant is 0.15 for all analyses in this paper. In tables 2 and 3, most of sectors are manufacturing. However, contrary to the intuitions, machinery and heavy industry do not play the important role. Because these sectors are the main players of international production networks, the effect of fragmentation will not be high in following empirical estimations. 4. Empirical methodology and hypotheses < insert table 2 > < insert table 3 > In this section, we try to examine the determinants of the two types of IIT in Japan using CEPII index. According to the theoretical implications reviewed in section 2, the determinants of vertical IIT and horizontal IIT may be different in Japanese IIT. Then, we estimate simple regressions to specify the important factors of Japanese vertical and horizontal IIT, and try to test that the theoretical implications are consistent with real economic activities. At first, for the vertical IIT, according to the theoretical implications, traditional determinants of trade such as factor contents or technological difference are mainly influenced to the trade amount. Because the information of the ratio of factor endowments is not available, we introduce the technological level as a variable for the decision factor of vertical IIT. For the variable of technological level, we use the R&D investment ratio per sales in each industry 9. We expect that the coefficient will be positive for the estimation of vertical IIT, and will show no correlation for horizontal IIT. Secondly, for the horizontal IIT, two variables are introduced. One is the variable for monopolistic competition. According to Lancaster (1980) and Helpman (1981), monopolistic competition is indispensable for horizontal IIT. Therefore, if the degree of market concentration is high, it is likely to increase horizontal IIT. For the variable of monopolistic competition, we introduce the natural log of average sales of firms in the industry in Japan. We expect that the coefficient of this variable have shows positive correlation in the estimation for horizontal IIT, and no correlation in the estimation for vertical IIT. Another is the variable for product differentiation. According to Krugman (1979, 1980) and Dixit and Norman (1980), the amount of horizontal IIT is likely to grow when the product variety is abundant because of the product differentiation, or 9 Typically, concentration ratio or Harfindahl index is used as a variable of monopolistic competition. However, our dataset do not include the information of the sales of each firm, and we cannot use these indices.

6 consumers prefer consuming many varieties of product. Because it is difficult to use the demand-side information or consumer preference, we use only the supply-side information such as product differentiation. For the variables of product differentiation, we introduce the advertising ratio per sales in the sector. This variable is expected to have positive coefficient in the regression on horizontal IIT, and not statistically significant coefficient in the regression on vertical IIT. In addition to these variables, the variables related to fragmentation theory are introduced. The existence of international production fragmentation increases vertical IIT because of the components or parts transaction. For the variables of fragmentation, average number of foreign subsidiaries in each industry and average number of East Asian subsidiaries in each industry are introduced. If a firm has foreign affiliate, it is expected to increase parts/components trade between parent firm and foreign affiliate. According to Kimura and Ando (2005), such production pattern is typical in East Asia. Then, we introduce the variable for subsidiaries in East Asian regions. Our estimation model is as follows: IIT=c+ (R&D)+ (Y)+ (ADV)+ (FAJF)+ (AAJF)+. (4) IIT is the index for vertical IIT or horizontal IIT measured as CEPII Index. R&D is R&D investment ratio per sales in industry and Y is the natural log form of average sales of a firm in the industry. ADV is the rate of advertising expenditure per total sales in industry. FAJF and AAJF are the average number of foreign affiliates of Japanese firms in the industry, and Asian affiliates of Japanese firms, respectively. IIT indices for vertical IIT and horizontal IIT is calculated to aggregate HS 6-digit commodities trade data derived from PC-TAS (UNCTAD) in 2000 to 2-digit industry. All information for other variables is derived from the Basic Survey of Japanese Business Structure and Activities (hereafter BSJBSA) (METI, GOJ). The summary and definition of all variables are shown in table Estimation results < insert table 4 > The results of estimation are presented at table 5. < insert table 5 > At first, R&D ratio (R&D) has positive correlation with Vertical IIT. This means that technological differences influence to the amount of vertical IIT, in other words, traditional determinants to trade is still effective to determine the amount of vertical IIT. This result is consistent with the theoretical implication reviewed in section 2, and implies the

7 theoretical model for explaining vertical IIT, for instance, quality ladder model, can explain the structure of Japanese trade pattern. On the other hand, because the variable does not show any correlation, horizontal IIT cannot be explained by the extension of traditional trade theories. Secondly, log of average sales (Y) has the positive and statistically significant coefficient only in the estimation for horizontal IIT. This result means that monopolistic competition is important for horizontal IIT but not important for vertical IIT. In other words, ideal variety model, such as Lancaster (1980) and Helpman (1981) which focused on the importance of monopolistic competition, is consistent with Japanese IIT pattern. On the other hand, advertising expenditure does not show the significant correlation with both IIT indices. Because the coefficient of this variable is expected the positive sign in horizontal IIT regression, this result is not consistent with the theoretical implication. It implies that the love of variety model, such as Krugman (1979, 1980) and Dixit and Norman (1980), has very limited power to explain Japanese IIT structure 10. Average number of foreign affiliates (FAJF) and average number of Asian affiliates (AAJF), introduced to estimate the impact of production fragmentation, have no correlation with both IIT indices. Though fragmentation is considered as one of the main factor of vertical IIT, we cannot find the importance of fragmentation in Japanese IIT 11. To summarize these results, we find that R&D ratio is important determinants of vertical IIT and average sales is important form horizontal IIT. These results imply that the extended model of traditional trade theory explains Japanese vertical IIT and monopolistic competition model, such as ideal variety model, explains Japanese horizontal IIT. Moreover, the results clarify that the determinants is solely different between vertical and horizontal IIT. 6. Conclusion In this paper, we analyze the determinants the vertical and horizontal IIT between Japan and rest of the world. According to various theoretical models, the determinants of vertical IIT and horizontal IIT are not same. Our empirical estimations confirm the theoretical implications empirically in Japanese IIT. Concerning with Japanese IIT patterns, vertical IIT is determined by the factors of traditional trade theory, such as technological differences, and horizontal IIT is influenced by the monopolistic competition, that is focused by ideal quality model. These results show various theoretical models can explain the phenomenon of real world in some extent, and it is important to discriminate vertical IIT with horizontal IIT. For further extension, we shall analyze the bilateral flow of IIT in spite of the trade with rest of the world in order to control the differences of factor contents or 10 However, many studies reported that advertising expenditure is not good proxy for the product differentiation of Japanese industry. Then, we should interpret this result carefully. 11 However, this result may come from the poor quality of the variables related to fragmentation.

8 technological differences between two countries. Reference Abd-el-Rahman, Kamal (1991), Firms Competitive and National Comparative Advantage as Joint Determinants of Trade Composition, Weltwirtschaftliches Archiv 127.1, pp83-97 Ando, Mitsuyo (2006), Fragmentation and Vertical Intra-Industry Trade in East Asia, forthcoming on North American Journal of Economics and Finance Arndt, Sven W. and Henryk Kierzkowski (2001), Fragmentation: New Production Patterns in the World Economy, Oxford University Press Aturupane, Chonira, Simeon Djankov and Bernard Hoekman, Horizontal and Vertical Intra-Industry Trade Between Eastern Europe and the European Union, Weltwirtschaftliches Archiv 135.1, pp62-81 Blanes, Jose V. and Carmela Martin (2000), The Nature and Causes of Intra-Industry Trade: Back to the Comparative Advantage Explanation? The Case of Spain, Weltwirtschaftliches Archiv 136.3, pp Brander, James A. (1981), Intra-Industry Trade in Identical Commodities, Journal of International Economics 11.1, pp1-14 Brander, James A. and Paul R. Krugman (1983), A Reciprocal Dumping Model of International Trade, Journal of International Economics 15.2, pp Cheng, Leonard K. and Henryk Kierzkowski (2001), Global Production and Trade in East Asia, Kluwer Academic Publishers Crespo, Nuno and Maria Paula Fontoura (2004), Intra-Industry Trade by Types: What Can We Learn from Portuguese Data?, Weltwirtschaftliches Archiv 140.1, pp52-79 Dixit, Avinash and Victor Norman (1980), Theory of International Trade: a Dual, General Equilibrium Approach, Cambridge; Cambridge University Press Falvey, Rod E. (1981), Commercial Policy and Intra-Industry Trade, Journal of International Economics 11.4, pp Finger, J. Michael (1975), Trade Overlap and Intra-Industry Trade, Economic Inquiry 13.4, pp Fontagne, Lionel and Michael Freudenberg (1997), Intra-Industry Trade: Methodological Issue Reconsidered, CEPII Working Paper Fontagne, Lionel, Michael Freudenberg and Nicholas Peridy (1997), Trade Patterns inside the Single Market, CEPII Working Paper Fukao, Kyoji, Hikaru Ishido and Keiko Ito (2003), Vertical Intra-Industry Trade and Foreign Direct Investment in East Asia, Journal of the Japanese and International Economies 17.4, pp Greenaway, David, Robert C. Hine and Chris Milner (1994), Country-Specific Factors and the Pattern of Horizontal and Vertical Intra-Industry Trade in the UK, Weltwirtschaftliches Archiv 130.1, pp Greenaway, David, Robert C. Hine and Chris Milner, Vertical and Horizontal

9 Intra-Industry Trade: A Cross-Industry Analysis for the United Kingdom, Economic Journal , pp Greenaway, David, Chris Milner and Robert J.R. Elliott (1999), UK Intra-Industry Trade with the EU North and South, Oxford Bulletin of Economics and Statistics 61.3, pp Grubel, Herbert G. and Peter J. Lloyd (1975), Intra-Industry Trade: The Theory and Measurement of International Trade in Differentiated Products, London; Macmillan Helpman, Elhanan (1981), International Trade in the Presence of Product Differentiation, Economies of Scale and Monopolistic Competition, Journal of International Economics 11.3, pp Hu, Xiaoling and Yue Ma (1999), International Intra-Industry Trade of China, Weltwirtschaftliches Archiv 135.1, pp Ishida, Osamu (2002), Nihon no Sangyonaiboeki, Suichokutekiboeki no Kozo (The Structure of Japanese Intra-Industry Trade and Vertical Trade), in Osamu Ishida and Hiroshi Fukagawa eds., Kokusai Boeki no Gurobaruka to Tayoka (2): Ajia Keizai to Gurobaruka (Globalization and Diversification in International Trade (2): Globalization and the Asian Economy), Kyushu Daigaku Shuppannkai (Kyushu University Press), pp1-26 Jones, Ronald W. (1956), Factor Proportions and the Heckscher-Ohlin Theorem, Review of Economic Studies 24.1, pp1-10 Kimura, Fukunari and Mitsuyo Ando (2005), Two-dimensional Fragmentation in East Asia: Conceptual Framework and Empirics, International Review of Economics and Finance 14.3, pp Krugman, Paul R. (1979), Increasing Returns, Monopolistic Competition and International Trade, Journal of International Economic 9.4, pp Krugman, Paul R. (1980), Scale Economies, Product Differentiation and the Pattern of Trade, American Economic Review 70.4, pp Lancaster, Kevin (1980), Intra-Industry Trade under Perfect Monopolistic Competition, Journal of International Economics 10.2, pp Lipsey, Robert E. (1976), Review of Grubel, H.G. and Lloyd, P.J., Intra-Industry Trade (1975), Journal of International Economics 6.2, pp Martin-Montaner, Joan A. and Vicente Orts Rios (2002), Vertical Specialization and Intra-Industry Trade: The Role of Factor Endowments, Weltwirtschaftliches Archiv 138.2, pp Mora, Carmen Diaz (2002), The Role of Comparative Advantage in Trade within Industries: A Panel Data Approach for the European Union, Weltwirtschaftliches Archiv 138.2, pp Rayment, Paul B.W. (1976), The Homogeneity of Manufacturing Industries with Respect to Factor Intensity: The Case of the UK, Oxford Bulletin of Economics and Statistics 38.2, pp Stiglitz, Joseph E. (1987), The Causes and Consequences of the Dependence of Quality on Price, Journal of Economic Literature 25.1, pp1-48 Veeramani, Choorikkad (2002), Intra-Industry Trade of India: Trends and

10 Country-Specific Factors, Weltwirtschaftliches Archiv 138.3, pp

11

12

13

14

15 ( 3 6 ) ( 0 0 )