Intra-Industry Trade Between Japan and Korea: Vertical Intra-Industry Trade, Fragmentation and Export Margins

Transcription

1 Intra-Industry Trade Between Japan and Korea: Vertical Intra-Industry Trade, Fragmentation and Export Margins Yushi Yoshida Faculty of Economics Kyushu Sangyo University This work is financially supported by Grant-in-Aid for Scientific Research (C)( ), JSPS. 1

2 The Objectives of This Paper This paper contributes to the existing empirical investigation of Japan-Korea international trade by providing new evidence of intra-industry trade between Korea and Japanese sub-regions. We calculate the Grubel-Lloyd intra-industry trade index for 41 regions of Japan with respect to Korea for the period between 1988 and The Grubel-Lloyd index at sub-regional level may capture the effect of the fragmentation of production. In intra-industry trade regression models, we introduce extensive and intensive margins of prefecture exports as new explanatory variables. 2

3 The Major Findings of This Paper By using Japanese prefecture international trade data, it is revealed that intra-industry trade is still pervasive even when it is restricted to trade flows between prefectures and Korea. We find that a rise in intra-industry trade is driven by the introduction of a new variety of exports, extensive margin, while intra-industry trade is discouraged by an increase in the trade value of products already exported, intensive margin. 3

4 The Structure of This Presentation (1)The Overview of International Trade between Japan and Korea (2)The Measurement of Intra-Industry Trade and Export Variety (3)Disaggregation by Prefectures (4)The Empirical Model and the Results (5)Discussion and Conclusion (6)Extensions and Other working projects 4

5 (1)The Overview of International Trade between Japan and Korea Korea is the third most important trade partner for Japan. However, in terms of (the number of) Japanese FDI inflows, Korea hosts less than expected. Still, the degree of intra-industry trade between Japan and Korea is very high. 5

6 (1-1)Trade Between Japan and Korea Table 1. Development of Japanese Trade with Korea Export Import (Trillion Yen) (percentage) (Trillion Yen) (percentage) Year Total Export Korea Total Import Korea (3) (5) (4) (4) (2) (3) (2) (3) (3) (3) (3) (4) Note: Total export(import) is the value of Japanese export(import) to the world. Figures in parenthesis are the rank of Korea as a trade partner in terms of trade values. (Source: author's calculation from the Japan Custom, the Ministry of Finance) 6

7 70 (1-2)Japanese FDI (the number of establishments) Figure 3. Japanese FDI in Korea and other Asian Economies (Number of New Establishments) 60 Singapore Taiwan Korea

8 (1-3)Intra-Industry Trade Figure 4. Intra-Industry Trade: Grubel-Lloyd indices in 1988 and Taiwan Philippine Korea Germany China Thailand Malaysia Guam Singapore UK France Sweden USA 0.15 Italy

9 (2)The Measurement of Intra-Industry Trade and Export Variety 9

10 (2-1)The Measurement of Intra-Industry Trade and Export Variety Example Foreign Country industry A Industry B (100) (40) (60) Home Country IIT ij The Grubel-Lloyd Index {(40 + 0) } + {( ) } = = = IIT K å ( X + X ) - X - X 2min( X, X ) ijk jik ijk jik ijk jik k = 1 k= 1 ij = = K K å X + X X + X ijk jik ijk jik k= 1 k = 1 K å å 10

11 (2-2)The Measurement of Intra-Industry Trade and Export Variety The determinants of IIT Similarity in GDP per capita (factor endowments) and larger size of economy (Helpman, 1987) Dissimilarity in GDP per capita via quality model ( Falm and Helpman,1987) and via fragmentation (Arndt and Kierzkowski, 2001) Disaggregation by sub-regions (Fragmentation) (1) (2) (3) (4) Region A Region B Region C Japan 11

12 (2-3) Recent papers flourishing on export margins Feenstra, R. and HL Kee, Export Variety and Country Productivity: Estimating the Monopolistic Competition Model with Endogenous Productivity, JIE, Arkolakis and Muendler, The Extensive Margin of Exporting Goods: A Firm-level Analysis, (2008, submitted paper for NASMES) Shepherd, B., Product Standards, Harmonization, and Trade: Evidence from the Extensive Margin, (2008 Midwest Spring Meeting). Besedes, T. and T. J. Prusa, The Role of Extensive and Intensive Margins and Export Growth, NBER WP No , Tarasov, A., Globalization: Intensive versus Extensive Margin, (2008, NA Summer Meeting ES). 12

13 (2-4)The Measurement of Intra-Industry Trade and Export Variety Extensive Margin and Intensive Margin, Hummels and Klenow (2005) Export Share = Extensive Margin * Intensive Margin EM ij = å å kîi ij kîi x x mjk mjk Value of world total export for the product set the country i exports. Value of world total export IM ij = å å kîi kîi ij ij x x ijk mjk Total export value for country i Value of world total export for the set Iij 13

14 (2-5)The Measurement of Intra-Industry Trade and Export Variety Extensive Margin and Intensive Margin, Yoshida (2008) Export Share = Extensive Margin * Intensive Margin EM ij = å å kîi ij kîi x x mjk mjk Value of Japanese total export for the product set the region i exports. Value of Japanese total export IM ij = å å kîi kîi ij ij x x ijk mjk Total export value for region i Value of Japanese total export for the set Iij 14

15 (2-6)The Measurement of Intra-Industry Trade and Export Variety Concentration and diversification of production location Case I Case II Region National Region National product A B C D sum product A B C D sum sum sum Case I (More specialized regions) Export Share = 0.25 Extensive Margin = 0.5 Intensive Margin = 0.5 Case II (More diversified regions) Export Share = 0.25 Extensive Margin = 0.75 Intensive Margin =

16 (3)Disaggregation by Prefectures IIT between Korea and Japanese prefectures Even when trade is broken down into the prefecture level, intra-industry trade still remains very high for some prefectures Extensive margins of Japanese prefectures with Korea Two types Consistently high EM prefectures Prefectures with growing EM 16

17 (3-1)Japanese Prefecture IIT for Korea Figure5. Selected Japanese Prefecture IIT for Korea Japan Chiba Fukuoka Osaka Okinawa Tokyo Hyogo Kanagawa Aichi Kagawa Nagasaki 17

18 (3-2)Japanese Prefecture Extensive Margins Figure 6. Extensive Margins of Selected Prefectures for Exports to Korea Osaka Kanagawa Chiba Hyogo Aichi Fukuoka Tokyo Yamaguchi Kyoto Hiroshima

19 (4)The Empirical Model and the Results 19

20 (4-1)The Empirical Model and the Results The traditional model IIT = a + b GDP _ KOR + b GDP _ PREF + b DGDPPC + e it i 1 t 2 it 3 it it The model with export margins IIT = a + b GDP _ KOR + b GDP _ PREF + b DGDPPC it i 1 t 2 it 3 it The Data + b EXTM + b INTM + b DIST + e 4 it 5 it 6 it The World Development Indicator, the Annual report on Prefectural Accounts, the Census Population Variable transformation Logistic transformation / Box-Cox transformation 20

21 (4-2)The Empirical Model and the Results The Effect of Intensive Margin on Intra-Industry Trade Helpman(1987) s : the share of the home country in world spending n : the number of differentiated product varieties IIT ij = sn * * s n An increase in n is an increase in within HS9 category. This should be interpreted as an increase in Intensive Margin. INTM IIT 21

22 (4-3)The Empirical Model and the Results The effect of Extensive margin on IIT New export Into an industry Which had no trade in both direction negative effect Which had some imports before positive effect Which is matched with new imports positive effect Our conjecture: A new export is created in the sector which is likely to have both direction of trade (IIT). EXTM IIT 22

23 Table2. Regression Results Model 1 Model 2 Model 3 Fixed Random Fixed Random Random GDP_KOR 2.216*** 2.080*** 0.695** 0.719*** 0.746*** (0.288) (0.288) (0.281) (0.267) (0.266) GDP_PREF (5.430) (2.115) (4.742) (1.764) (1.748) DGDPPC (5.171) (1.168) (4.482) (0.970) (0.967) EXTM 1.278*** 1.243*** 1.204*** (0.157) (0.100) (0.102) INTM *** *** (0.184) (0.000) (0.109) DIST (0.967) Observations No. of prefectures adj. R Hausman: CHISQ p-value Note: The dependent variable is the Box-Cox logistic transformed Grubel-Lloyd index. The extensive margin (EXTM) and intensive margin (INTM) are Box-Cox transformed. The parameter lamda for Box- Cox is set equal to 0.1. The other explanatory variables are in logarithmic form. Figures in parenthesis are standard errors (heteroskedasticity-consistent for fixed model). The Hausman statistics, given as CHISQ, tests the null of consistency of random effect estimates. Statistical significance at one, five, and ten percent are indicated by ***, **, and *, respectively. 23

24 (5)Discussions and Conclusions 24

25 (5-1)Discussions and Conclusions We observed that Korea had become one of the highest intra-industry trade partners of Japan by Even when disaggregated to prefecture levels, a high degree of intra-industry trade persisted among many regions. Since we are restricting the individual area to be 2.5% (on average) of the nation, intra-industry trade is driven by the fragmentation of production between Korea and Japan. 25

26 (5-2)Discussions and Conclusions Two new hypotheses: (1) INTM IIT that intra-industry trade between Japanese prefectures and Korea may be lowered by raising the intensity of trade for the products a prefecture already traded, given that the prefecture is the net exporter of the differentiated products (2) EXTM IIT that intra-industry can be strengthened by engaging in new trade for the products if matched by imports. We confirmed our hypotheses by obtaining significant coefficients with correct signs for both extensive and intensive margins. 26

27 (6)Extensions and Preliminary Results 27

28 (6-1)Extensions and Preliminary Results We extended partner countries of Japan to include other eight Asian economies. (China, Taiwan, Hong Kong, Singapore, Malaysia, Indonesia, Philippines, and Thailand) IIT it = a + i b + b 4 1 GDP _ KORt + b 2GDP _ EXTM it + b INTM 5 it + b DIST 6 PREF it + e it + b DGDPPC 3 IIT = a + b MaxGDP + b MinGDP + b DGDPPC ijt i 1 ijt 2 ijt 3 ijt + b EXTM + b INTM + b DIST + e 4 ijt 5 ijt 6 ij ijt 28

29 (6-2)Extensions and Preliminary Results Dependent variable: IIT Adjusted R-squared = Number of observation = 6129 Variable Coeff. P-value C [.000] MAXGDP [.000] MINGDP [.000] D_PC [.000] EXTM [.000] INTM [.586] DIST [.038] (1) Positive estimated coefficient for difference in GDP per capita. Fragmentation not product variety (2) Negative coefficient for distance. Larger IIT with closer countries 29

30 (6-3)Present and Future Works Currently working on: 1) Gravity Model between Japanese Ports and Asian Regions Export-Platform Domestic Investments with Proximity and Product Differentiation: Empirical Evidence from Port-level Analysis 2) Exchange Rate Pass-through by Japanese Ports A New Evidence for Exchange Rate Pass-through: Disaggregated Trade Data from Local Ports 3) Export Margins of Japanese Prefectures Industry Location and Variety Growth: Empirical Examination of Regional Exports 4) Intra-Industry Trade Japan-Korea, Japan-Taiwan, Japan-Asia Future Works (5) Industry-Region Analysis with trade and production (with Chul Chung) 30

31 Thank you!!! 31