Foreword This volume is a collection of papers which were presented at the Conference "The EU and Emerging Markets" org...
8 downloads
565 Views
1MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Foreword This volume is a collection of papers which were presented at the Conference "The EU and Emerging Markets" organized by the Research Institute for European Affairs and ECSA Austria at Vienna University of Economics and Business Administration on 5/6 June 2007. The intention of this conference was to discuss the manifold aspects which evolve in the relation of the EU with emerging markets in Eastern Europe, Asia and Latin America both with trade and investment. These markets have become more and more important for the EU which is reflected in the EU´s important initiatives for trade agreements with these economies. The volume starts with the papers presented by Dominick Salvatore, Fordham University New York, and Wilhelm Kohler, University of Tübingen, in their keynote lectures. Dominick Salvatore discusses the latest developments of the EU´s trade with emerging markets and shows that the EU can defend a comparative advantage vis-à-vis all its trading partners except for the Asian economies. Wilhelm Kohler presents a rigorous theoretical model which explains the activities of offshoring, i.e. the fragmentation and relocation of stages of production, by factor cost advantages and draws conclusions concerning skills, wages and employment in the domestic country. Part 2 contains three papers which focus on the EU´s relation with Eastern Europe. Martin Falk and Yvonne Wolfmayr analyse the employment and turnover effects of Austrian and German firms that invest in affiliates in Eastern Europe and conclude that investors show higher growth effects compared to non-investing domestic companies. Özlem Onaran presents an empirical investigation on the employment effects in the integrating Eastern European economies and finds that the expected employment effects are only weak. Ledyaeva provides an empirical analysis of regional FDI inflows in Russia which turns out to be concentrated on big cities and regions rich in oil. Part 3 focuses on the relation of the EU with Latin American emerging markets. Inmaculada Martínez-Zarzoso investigates the impact of transport costs, infrastructure and location on LA imports from the EU finding that improvements would increase trade by up to 30 per cent which is particularly relevant in standard product categories. Rodrigo Zeidan analyses trade patterns between Brazil
VI
and the EU and identifies a typical North-South trade structure which has remained fairly stable over the past decade. Maria Bas and Ivan Ledezma wish to assess to which extent Chile has achieved trade integration. They find that Chile has more liberalized towards imports than have its trading partners and that liberalization is very sector specific. The paper of MartínezZarzoso, Nowak-Lehmann and Vollmer verifies the dominance of Ricardian trade theory in trade patterns of European countries with developing countries, among them Brazil. They examine industry level unit value costs. Part 4 looks at the relations of the EU with Asian economies. Bernadette Andreosso-O´Callaghan and Françoise Nicolas question the complementarity of trade relations between the EU and the ASEAN economies in view of the intensified dialogue for trade liberalization between the two parties. They find that the EU has gained an important position in the ASEAN markets whereas for those economies trade relations with other Asian countries dominate by far. The second paper of this part by Hien Minh Dinh et al. describes the impressive export performance which Vietnam could watch during the past ten years. Their analysis shows that this is not only the effect of increasing world demand but also of improved competitiveness of Vietnam. During the conference fruitful discussions evolved among the participants and it became clear that the EU is watching a big challenge by new trade patterns with emerging markets. The editor appreciates that almost all conference participants could contribute to this volume. I am indebted to Roman Puff for his assistance in organizing the conference and above all to Nurgül Özen for her secretarial work in elaborating the manuscripts. Vienna, March 2009
Gabriele Tondl
Table of Contents
Part 1:
Introduction: The EU and Emerging Markets
1
Keynote Lecture 1: Dominick Salvatore European Union Trade and Investment Relations with Emerging Markets
3
Keynote Lecture 2: Wilhelm Kohler Offshoring: Why Do Stories Differ?
17
Part 2:
51
The EU and Eastern Europe
Martin Falk and Yvonne Wolfmayr Home Market Effects of Outward FDI: Evidence Based on Amadeus Firm-Level Data
53
Özlem Onaran The Effect of Trade and FDI on Employment in Central and Eastern European Countries: A Country-Specific Panel Data Analysis for the Manufacturing Industry
71
Svetlana Ledyaeva Foreign Direct Investment Determinants in Post-Crisis Russia: Empirical Evidence
95
Part 3:
The EU and Latin America
109
Inmaculada Martinez-Zarzoso On Transport Costs and Sectoral Trade: Further Evidence for Latin-American Imports from the European Union
111
Rodrigo M. Zeidan Evolving Patterns of Trade between Brazil and the European Union 1994-2005
133
VIII
Maria Bas and Ivan Ledezma Trade Integration in Manufacturing: The Chilean Experience
167
Inmaculada Martinez-Zarzoso, Felicitas Nowak-Lehmann D. and Sebastian Vollmer Comparing French and Spanish Exports to Emerging and Developed Markets
187
Part 4:
203
The EU and Asian Economies
Bernadette Andreosso-O´Callaghan and Françoise Nicolas Assessing the Complementarity between the EU and ASEAN Economies
205
Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh Vietnam’s Export to the EU: An Overview and Assessment using the Constant Market Share based Approach
235
List of Authors
263
Part 1: Introduction: The EU and Emerging Markets
Dominick Salvatore
European Union Trade and Investment Relations with Emerging Markets Abstract Emerging market economies are an important trade and investment partner for the European Union (EU). We measure the EU comparative advantage in manufactured goods and high-technology products using first the actual exchange and growth rates, and then an approximate equilibrium exchange rate and potential growth rate. It turns out that the EU has a comparative advantage vis-à-vis all its trading partners except for the Asian economies. Subsequently, the paper examines EU investments in emerging markets which account for 33 per cent of all EU outward FDI. Finally, we show that the EU has a number of trade agreements with emerging market economies in effect or under negation. I. Introduction The European Union (EU) trade and investment relations with emerging markets are now dominated by three major issues: 1. The emergence of China (and to a lesser extent India) as major trade competitors on the world scene. 2. The negotiations of (and collapse and attempts to revive) the Doha Round of multilateral trade negotiations, and 3. The alleged double-competitive squeeze that the EU has faced during the last decade or so and loss of international competitiveness, in high-tech goods vis-à-vis the United States and Japan, and in other manufactured goods vis-à-vis emerging markets, especially China. We will see that some of the above views and concerns are valid while others are not. But to determine this, we start by examining data on EU trade and investment with emerging markets and other advanced nations in today’s world context. The paper starts presenting data on overall EU merchandise exports and measures its comparative advantage in manufactured goods and high-technology products. Subsequently, it re-estimates
4
European Union Trade and Investment Relations with Emerging Markets
the EU comparative advantage using an approximate equilibrium exchange rate and the potential EU growth rate. Finally, the paper examines EU investments in emerging markets and its trade and investment policies toward these economies. II. Destination of EU merchandise exports Emerging markets have become an important destination of EU exports. Table 1 shows the value of EU-25 merchandise exports to the world (including, and excluding, intra-EU-25 exports) and to various regions and main trading partners in 2001, 2003, 2005, as well as their growth from 2001 to 2005. From the table we see that EU-25 exports to the world amounted to $4,000 billion in 2005, if intra-EU-25 exports are included, and $1,327 billion if intra-EU-25 exports are excluded. However, since the EU is a single economic block, the latter figure is in fact relevant (note that the United States does not count the trade between California and New York State as international trade). The $1,327 billion of EU-25 exports to the rest of the world represents about 10 per cent of its GDP (if intra EU-25 exports were included the export rate would be 30 per cent of GDP). Table 1: EU-25 merchandise exports by destination (billions of U.S. Dollars, 2001-2005) Destination 2001 World (incl. intra EU-25) 2,452 World(excl. intra EU-25) Europe Extra EU-25 North America - USA Asia - Japan - China - Six East Asian CIS - Russian Federation South-Central America - Brazil Africa - S. Africa Middle East Source: WTO, 2006.
802 141 252 217 182 40 27 72 39 28 42 16 62 11 58
2003 3,124 999 190 296 254 231 45 45 83 59 41 40 14 78 15 75
2005 Growth % 2001-2005 4,000 13 1,327 263 360 309 304 53 64 95 99 70 54 20 105 22 107
13 13 9 9 14 7 24 7 26 26 7 5 14 19 17
Dominick Salvatore
5
Table 1 also shows that of total EU-25 exports of $1,327 billion (excluding intra EU-25 exports) in 2005, $263 billion went to other European countries, $360 billion went to North America (of which $309 billion to the United States), $304 billion to Asia (of which, $53 billion to Japan, $64 billion to China, and $95 billion to the six other dynamic emerging market economies in Asia – Malaysia, Singapore, South Korea, Taiwan, Thailand, and the Philippines), $99 billion to the Commonwealth of Independent States CIS (of which, $70 billion to the Russian Federation), $54 billion to South and Central America (of which, $20 billion to Brazil), $105 billion to Africa (of which, $22 billion to South Africa), and $107 billion to the Middle East. As indicated in the last column of Table 1, the average growth of EU-25 exports over the 2001-2005 period amounted to 13 per cent vis-à-vis the rest of the world, 9 per cent with respect to North America and the United States, 14 per cent with Asia (24 per cent with China), 26 per cent vis-à-vis CIS and Russia, only 7 per cent with South and Central America, 14 per cent with Africa (19 per cent with South Africa), and 17 per cent with the Middle East. Thus over the 2001-2005 period EU-25 exports grew fastest with Russia, followed by China and South Africa. III.
EU-25 comparative advantage in manufactured goods
The EU seems to be more competitive in its export products than generally considered. Table 2 measures the revealed comparative advantage or disadvantage in all manufactured goods of EU-25 in 2005, overall or in total, and vis-à-vis other major regions and countries. The revealed comparative advantage is measured by dividing the net trade balance (exports minus imports) of EU-25 with respect to other regions and countries by total exports of manufactured goods of EU-25. A positive sign indicates a comparative advantage and a negative sign a comparative disadvantage, while the absolute value of the calculated index shows the extent of the comparative advantage or disadvantage. For example, the overall extra EU-25 net trade balance in all manufactured goods in 2005 was $198.11 billion ($1,123.59 of exports minus $925.48 of imports). Dividing the extra EU-25 net trade balance in manufactured goods of $198.1 billion by total extra EU- 25 manufactured exports of $3,240.27 billion gives the index
6
European Union Trade and Investment Relations with Emerging Markets
value of 6.11, indicating a very strong overall comparative advantage in manufactured goods of EU-25 in 2005. Table 2: EU-25 trade and comparative advantage in manufactured goods vis-à-vis its main trading partners (billions of dollars, 2005) Trading partner
Exports
Europe Extra 1,123.59 EU-25 United States 262.31 North America 43.55 minus US Japan 49.17 56.56 China 162.54 Asia minus Japan & China Russian 61.06 Federation CIS minus 26.06 Russian Fed. Brazil 16.49 32.03 S-C America minus Brazil Africa (no data 86.69 S. Africa) Middle East 93.93 Source: WTO, 2006.
Imports
Net Bal.
925.48
198.11
Comp. Adv. (+) or Disadv. (-) 6.11
171.89 21.35
90.42 22.20
2.79 0.69
82.21 187.60 198.80
-36.24 -131.04 -36.26
-1.01 -4.04 -1.12
15.20
45.86
1.42
8.74
17.32
0.53
9.99 11.92
6.50 20.11
0.20 0.62
37.33
49.36
1.52
26.18
67.75
2.09
The last column of Table 2 indicates that in 2005 EU-25 had a strong (revealed) comparative advantage in overall manufactured goods vis-à-vis the United States (2.79). This is opposite to the general popular belief which is related to the fact that the United States consistently ranked much higher than Europe on international competitiveness scores during the past decade. EU-25 also seems to have a strong comparative advantage in manufactured goods with respect to the Middle East (2.09), Africa (1.52), and Russia (1.42), and less vis-à-vis North America minus the United States (0.69), South and Central America minus Brazil (62), CIS minus la Russia (0.53), and Brazil (0.20). On the other hand, EU-25 has a very strong comparative disadvantage in manufactured goods
Dominick Salvatore
7
with respect to China (-4.04), and to a lesser degree with respect to Asia minus China and Japan (-1.12), and Japan (-1.01). IV.
EU-25 comparative advantage in high-tech products
The EU is also competitive in high-tech products. Table 3 measures the revealed comparative advantage or disadvantage in high-tech products of EU-25 in 2005, overall, and vis-à-vis other major regions and countries. High-tech products include chemicals (which includes pharmaceuticals), machinery (which includes office and telecommunications equipment), and transport equipment (which includes automobiles, which despite being a mature product, incorporate a great deal of high technology). The last column of Table 3 indicates that in 2005 the EU-25 had a very strong (revealed) comparative advantage in high-tech products, with an overall index of 5.85, and with respect to the United States (1.56), Middle East (1.51), and Africa (1.50), and with an index of less than one with the other areas and countries. At the same time, EU-25 had a relative strong comparative disadvantage in high-tech products vis-à-vis Japan (-1.29), China (-1.19), and Asia minus Japan and China (-0.91). Table 3: EU-25 trade and comparative advantage in high-tech products vis-à-vis its main trading partners (billions of dollars, 2005) Trading partner Europe Extra EU-25 United States North America minus US Japan China Asia minus Japan & China Russian Federation CIS minus Russian Fed. Brazil S-C America minus Brazil Africa Middle East Source: WTO, 2006.
Exports
Imports
Net Bal.
747.39 187.80 39.55
557.83 137.16 11.02
189.56 50.64 28.53
Comp. Adv. (+) or Disadv. (-) 5.85 1.56 0.88
33.42 44.06 112.99
72.36 82.77 142.51
-38.94 -38.71 -29.52
-1.20 -1.19 -0.91
36.65 24.95
5.98 2.66
30.67 22.29
0.94 0.69
13.08 25.30
5.77 8.43
7.31 16.87
0.23 0.52
63.51 66.44
14.81 17.66
48.70 48.78
1.50 1.51
8
European Union Trade and Investment Relations with Emerging Markets
V. EU-25 comparative advantage in office and telecommunication equipment Table 4 measures the revealed comparative advantage and disadvantage in office and telecommunications equipment of EU-25 in 2005, overall and vis-à-vis other major regions and countries. Office and telecommunications equipment is, of course, part of high-tech products, but we discuss their revealed comparative advantage separately because it is one of the most dynamic sectors of advanced economies and also because of their importance in determining the international competitiveness of a nation or region. The last column of Table 4 indicates that in 2005 EU-25 had a very strong (revealed) comparative disadvantage in office and telecommunications equipment, with an overall index of -2.56, and with China (-1.73) and Asia Minus Japan and China (-1.47), and much less with Japan (-0.51), the United States (-0.20), and North America minus the United States (-0.05), but a weak comparative advantage with other major regions and countries. Table 4: EU-25 trade and comparative advantage in office and telecom equipment vis-à-vis its main trading partners (billions of dollars, 2005) Trading Partner
Exports
Imports
Net Bal. Comp. Adv. (+) or Disadv. (-) -82.89 -2.56 -6.32 -0.20 -1.55 -0.05
Europe Extra EU-25 United States North America minus US Japan China Asia minus Japan & China Russian Federation CIS minus Russian Fed. Brazil S-C America minus Brazil Africa Middle East Source: WTO, 2006.
118.28 17.87 2.01
201.17 24.19 3.56
2.84 5.32 25.23
19.49 61.22 72.82
-16.65 -55.90 -47.59
-0.51 -1.73 -1.47
9.87 3.20 1.04 12.03
0.09 0.16 0.39 -0.14
8.78 3.04 0.65 12.17
0.27 0.09 0.02 0.38
9.57 17.49
0.90 6.97
8.67 10.52
0.27 0.32
The overall conclusion is that EU-25 has a very strong comparative advantage in manufactured goods, as a whole, and in high-
Dominick Salvatore
9
tech products with respect to the United States and the rest of the world, excluding Asia. In office and telecommunication equipment, EU-25 has a small comparative disadvantage with respect to other developed countries, but a strong one with respect to China and the other Asian tigers (Asia minus Japan and China). Thus, in general, EU-25 has a comparative advantage in manufactured goods with respect to the United States and comparative disadvantage in hightech products with Asia, and even with the United States. VI. EU-25 revealed comparative advantage and disadvantage in manufactured goods with an equilibrium exchange rate and potential EU growth Table 5 re-estimates the EU-25 comparative advantage and disadvantage in all manufactured goods using an estimated equilibrium euro/dollar exchange rate and assuming that EU-25 achieved its potential growth rate in 2005. The reason is that an overvalued euro leads to an underestimation of the EU-25 comparative advantage and to an overestimation of its comparative disadvantage. The opposite is the case with an EU-25 growth rate below its potential, as it was the case in the year 2005. In 2004 the average dollar/euro exchange rates was $1.24, which made the euro overvalued by 7.8 per cent with respect to the dollar, if we assume that the equilibrium exchange rate is $1.15. With a short-run price elasticity of EU-25 exports of (-)0.175 and short-run price elasticity of imports of (-)0.075 (see Hooper and Marquez, 2000), this increases the EU-25 comparative advantage and reduces its comparative disadvantage in overall manufactured goods, as indicated in the column headed “Short-Run”. For example, the European extra EU-25 comparative advantage in manufactured goods as a whole increases from 6.11 to 6.76 (or by about 11 per cent) if the euro exchange rate had been at the equilibrium rate of $1.15 (instead of being overvalued by 7.8 per cent at the actual exchange rate of $1.24 in 2005). The rest of the column indicates the extent to which the EU-25 comparative advantage would be higher and comparative disadvantage lower in manufactured goods as whole with respect to various countries and regions in 2005. In general, the data indicates that the EU-25 comparative advantage is higher (and the EU discomparative advantage lower), on average, by 1-5 per cent.
10
European Union Trade and Investment Relations with Emerging Markets
The column headed “Long-Run” in Table 5 provides the result of re-estimating the EU-25 comparative advantage and disadvantage in manufactured goods using the long-run price elasticity of exports of (-)0.75 and long-run price elasticity of imports of (-)0.38. This leads to a higher European extra EU-25 comparative advantage in manufactured goods as a whole from 6.11 to 8.99 (or by about 47 per cent) if the euro exchange rate had been at the equilibrium rate of $1.15 instead of $1.24). The rest of the column indicates the extent to which the EU-25 comparative advantage would be higher and comparative disadvantage lower in manufactured goods with respect to various countries and regions in 2005. In general, the data indicates that the EU-25 comparative advantage increases (and the EU comparative disadvantage decreases) on average by 10-20 per cent. The last column of Table 5 shows the re-estimated comparative advantage and disadvantage of EU-25 using the equilibrium exchange rate of $1.15 and assuming that EU-25 grew at its average potential growth rate of 2.25 per cent, instead of the actual rate of 1.50 per cent. Since at the higher potential growth rate, EU-25 imports would be higher, its revealed long-term comparative advantage would be somewhat lower and its comparative disadvantage somewhat higher than at the actual growth rate. Table 5: Comparative advantage in manufactured goods of EU-25 with an Equilibrium exchange rate and potential growth in 2005 Trading partner Europe Extra EUUnited North America Japan China Asia minus Russian CIS minus Russian Brazil S-C America Africa Middle
Actual
Short-Run
Long-Run
Long-Run and +1% Growth
6.11
6.76
8.99
8.71
2.79 0.69
2.93 0.71
3.42 0.78
3.37 0.78
-1.01 -4.04 -1.12
-0.98 -3.99 -1.01
-0.86 -3.77 -0.64
-0.88 -3.83 -0.70
1.42 0.53
1.44 0.55
1.54 0.59
1.53 0.59
0.20 0.62
0.21 0.64
0.24 0.69
0.24 0.69
1.52 2.09
1.57 2.14
1.71 2.28
1.70 2.28
Dominick Salvatore
11
Using an income elasticity of EU-25 imports of 1 (OECD, 2001) we get the values indicated in the last column of Table 5. It shows that the European extra EU-25 comparative advantage in manufactured goods as a whole falls from 8.99 with the actual EU-25 growth rate to 8.71 (or by only about 3 per cent) with the EU-25 growing at its potential rate in 2005. The rest of the column indicates the extent to which the EU-25 comparative advantage would be higher and comparative disadvantage lower in manufactured goods as a whole with respect to various countries and regions in 2005. In general, the data indicates that the EU-25 comparative advantage would be higher (and the EU comparative disadvantage lower), on average, by between 8-18 per cent. VII. EU stock and flows of FDI to emerging markets Emerging markets account for a large share of EU outward FDI. Table 7 shows the stock and flows of extra EU-25 FDI, in total, and to emerging market economies in 2003 and 2004. The table shows that the total extra EU-25 stock of FDI was $1,976 billion in 2003, of which $668 billion or 33.8 per cent went to emerging market economies. The individual emerging market economies with the highest stock of FDI from EU-25 in 2003 were China (with $93 billion, or 4.7 per cent of the total extra EU-25 stock of FDI), Brazil ($48 billion, with 2.4 per cent of the total), Singapore ($45 billion or 2.3 per cent), Mexico ($27 billion or 1.4 per cent), Argentina ($23 billion or 1.2 per cent), South Korea ($15 billion or 0.7 per cent), Russia ($13 billion or 0.7 per cent), and India ($7 billion or 0.3 per cent). The stock of FDI in Africa in 2004 was $79 billion or 4.0 per cent of the total, in Other Asia $37 billion or 1.9 per cent of the total, in the Candidate Countries (Bulgaria, Croatia, Romania, and Turkey) $25 billion or 1.3 per cent of the total, and in the Near & Middle East had $15 billion or 0.8 per cent of the total. The last column of Table 6 shows that the total extra net outflows of FDI from EU-25 in 2004 were $115 billion, of which $76 billion or 66.2 per cent went to emerging market economies. The individual emerging market economies with the highest flows of FDI from EU-25 in 2004 were China (with $11 billion, or 9.6 per cent of the total extra EU-25 flows of FDI), Mexico (with $9 billion or 7.6 per cent), Russia (with $6 billion or 5.2 per cent), Brazil and Singapore (with $3 billion each or 2.6 per cent), Korea and India
12
European Union Trade and Investment Relations with Emerging Markets
(with $1 billion or 0.9 per cent), while Argentina experienced an outflow of $2 billion. Africa received $12 billion or 10.4 per cent of total FDI from EU-25, Other Asia $9 billion or 7.8 per cent, and the Near & Middle East $1 billion or 0.9 per cent. Table 6: EU-25 stock and flows of FDI to emerging markets (billion euros, 2003-2004) stock 2003 1,976
% of total 100
Total extra EU-25 Total 668 33.8 Emerging Markets Russia 13 0.7 Candidate 25 1.3 Countries* Brazil 48 2.4 Mexico 27 1.4 Argentina 23 1.2 China 93 4.7 Singapore 45 2.3 S. Korea 15 0.7 India 7 0.3 Other Asia 37 1.9 Near & 15 0.8 Middle East Africa 79 4.0 * = Bulgaria, Croatia, Romania, and Turkey. Source: Eurostat, 2006.
flows 2004 115
% of total 100
76
66.2
6 4
5.2 3.5
3 9 -2 11 3 1 1 9 1
2.6 7.6 -1.7 9.6 2.6 0.9 0.9 7.8 0.9
12
10.4
VIII. EU trade and investment policies toward emerging markets The EU has attempted to negotiate free trade agreements with a number of emerging markets. Figure 1 shows the EU’s regional trade arrangements (RTA): those signed and in force (Chile, Mexico, South Africa, and Overseas Countries and Territories or OCT); those under negotiations (Mercosur and Gulf Cooperation Council (GCC); and those under consideration (India and Korea) as of 2007. There is a great deal of disagreements whether these RTA represent “building blocks” or “stumbling blocks” toward a liberal multilateral trading system.
Dominick Salvatore
13
The EU was, of course, a major player in the Doha negotiations with emerging markets during the 2001-2006 period. The collapse of those negotiations can be attributed mainly to the objections of advanced countries to liberalize trade in agricultural products, while at the same time requesting that emerging markets open up their industrial and financial sectors to freer trade and investments. The United States, of course, protects its agriculture as much as the EU but seems genuinely interested to freer trade in agricultural products, perhaps as a way to overcome the strong domestic agricultural lobby and so reduce the huge subsidies it provides to its farmers. The EU, and particularly France, supports agriculture so as to protect a rural way of life that is fast disappearing. But this should not be at the expense of the agricultural exports of developing countries, but perhaps by stressing specialty products, and organic foods, and hormone-free products. Figure 1: European Intra and Cross-Regional TRA Network
Mexico
Source: Crawford and Farentino, 2007.
14
European Union Trade and Investment Relations with Emerging Markets
Advanced countries, but especially the EU, want to impose strong labor and environmental standards on developing countries, supposedly to protect workers from exploitation and the environment from pollution in developing countries. There is a strong suspicion, however, that this is a way of advanced countries protecting its workers and industries from competition from developing countries. Developing countries point out that the imposition of strong labor and pollution standards would nullify a great deal of their comparative advantage, and that if they could afford these standards they would not be developing countries in the first place. They add that, as they develop and become richer, their population will demand better labor and pollution standards and their governments would then be able to respond and introduce more effective standards. The EU is now under great pressure from an overvalued Euro. This discourages its exports, encourages its imports, and greatly distorts its comparative advantage. The overvalued euro results from the fact that the dollar needs to depreciate in order to reduce the unsustainable trade deficit of the United States, but since China does not allow its currency to appreciate significantly (and so do other Asian countries fearing greater Chinese competition), the dollar depreciated very heavily with respect to the euro, even though the EU is not itself responsible for the huge US trade deficit. To overcome this situation, Europe has now joined the United States in demanding a more rapid and a stronger appreciation of the Chinese Yuan. IX. Conclusions The calculations presented in this paper indicate that, in general, EU-25 has a comparative advantage in manufactured goods with respect to the United States, and comparative disadvantage in hightech products with respect to Asia and even with respect to the United States. If the dollar/euro exchange rate had been in equilibrium instead of being overvalued, the EU-25 comparative advantage would have been higher and comparative disadvantage lower by 1-5 per cent in the short run, by 10-20 per cent in the long run, and by 8-10 per cent in the long run if the EU-25 had grown at its potential rate. The total Extra EU-25 stock of FDI was $1,976 billion in 2003, of which $668 billion or 33.8 per cent of the total was in emerging
Dominick Salvatore
15
market economies, with the lion’s share going to China. Total net outflows of FDI to extra EU-25 in 2004 amounted to $115 billion, of which $76 billion or 66.2 per cent went to emerging market economies, with more going to China than any other single country. The EU has regional trade agreements (RTA) with a number of developing countries and areas and is negotiating others. The EU was a major player in the Doha Round of multilateral trade negotiations, which collapsed primarily because of the refusal of advanced countries to open up their markets more widely to agricultural food exports from developing countries, while at the same time demanding stronger labor and environmental standards in developing countries. The grossly overvalued euro is now putting strong pressure on EU industries and exports and distorting its comparative advantage.
References
Crawford, J.A., Farentino, J. (2005), The Changing Landscape of Regional Trade Agreements, WTO Discussion Paper No. 8, WTO, Geneva. Eurostat (2007), Database. IMD (2007), The World Competitiveness Yearbook, IMD, Lausanne. Hooper, P., Johnston, K., Marquez, J. (2000), Trade Elasticites for the G-7 Countries, Princeton Studies in International Economics No. 87, 8-9. OECD (2001), Economic Outlook, OECD, Paris. Salvatore, D. (1998), Europe’s Competitiveness Problems, The World Economy 21 (2), 189-205. Salvatore, D. (2003), The New Economy and Growth in the G-7 Countries, Journal of Policy Modeling 25 (5), 531-540. WEF (2007), The Global Competitiveness Report, WTO, Geneva. WTO (2006), International Trade Statistics, WTO, Geneva.
Wilhelm Kohler*
Offshoring: Why Do Stories Differ? Abstract This paper identifies critical modelling choices, as well as differences in the driving forces behind offshoring, that may explain differences in results. Offshoring of industry-specific tasks has wage and employment effects that are vastly different from those identified in Grossman and Rossi-Hansberg (2006), depending on how the industries differ in their average and marginal skill-intensities, respectively. Structural adjustment may occur at the intensive margin and the extensive margin (offshoring), and it may occur in opposite directions or the same direction at both margins, again depending on how industries differ in terms of their average and marginal skill-intensity. I. Introduction It is now almost 20 years since Jones and Kierzkowski (1990) have first pointed out the emergence of a new trend in globalization that was made possible mainly because of advances in the technology of transportation and communication: the fragmentation of production across borders. In the 1990s, this phenomenon became known as outsourcing. Krugman (1995) was the first to argue that an increased tendency towards outsourcing could be a partial explanation for the rising skill-premium in wages observed to a varying degree in many advanced industrial countries. Feenstra and Hanson (1997) were the first to offer a theoretical model for this type of explanation, with the important property that outsourcing would lead to the same effect on the skill-premium in both, the North (where outsourcing takes place) and the South (the target country of such outsourcing activities). Subsequently, a sizable literature has devel*
I wish to acknowledge financial support received from Fritz Thyssen Foundation under grant no Az. 10.06.1.111. Thanks are due to Christoph Roth for thoughtful and critical remarks on an earlier version of the paper.
18
Wilhelm Kohler
oped, focusing either on the specific conditions that favor outsourcing, or on the effects that it has on the outsourcing country's wages, employment, and welfare. In this process the terminology has undergone certain changes, whereby outsourcing relates to procurement via market-based transactions versus in-house operation of a given stage of the production process. What was originally called international fragmentation or outsourcing is now commonly referred to as offshoring, meaning that firms delocalize certain production-stages to other countries in order to arbitrage on international factor-cost differences. Obviously this can either happen through an in-house operation, or through cross-border outsourcing. Empirical measurement of offshoring notoriously suffers from certain problems of precise definition and data limits, but it is probably fair to conclude from existing studies that the quantitative significance of offshoring is relatively low, at least below what would justify the enormous attention that offshoring has drawn in the policy debate; see for instance Bhagwati et al. (2004). More recently, this attention has even increased, due to evidence that the practice of offshoring is by no means restricted to production stages intensive in low-skilled labor and requiring relatively little highskilled labor. Several authors have been arguing that much of the additional offshoring to be expected for the future will probably affect high-skilled workers much more than was the case in the past, maybe even more than low-skilled labor; see for instance Markusen (2006) and Blinder (2006). From existing empirical studies, it is relatively easy to quote both, numbers that portray offshoring as a relatively minor phenomenon, and numbers suggesting that it marks a major shift in international trade. In a recent study, the OECD calculates a countryspecific index of offshoring, based on the share of imported nonenergy intermediate inputs in various industries' total use of nonenergy intermediates. On this account, the smaller European countries, like Belgium, Austria and Denmark, reveal a high level of offshoring, with index-values for 2000 in the vicinity of 0.7.1 Large countries like Japan and the US, however, appear much less prone to offshoring, with index-values well below 0.2. In all countries considered, the offshoring-index for 2000 is larger than 1
An index value of 0.7 means that on average a country's industries are estimated to rely on imports for 70 percent of their non-energy intermediate inputs; see OECD (2007, pp. 61).
Offshoring: Why Do Stories Differ?
19
for 1995, although growth rates rarely exceed 20 per cent. Interestingly, the index values are largest for manufacturing intermediates in manufacturing industries, about 7 to 8 times the values for service intermediates and service industries. However, such indices have very limited information content. They certainly cover much more than would be considered as offshoring in theoretical analyses or in the policy debate. Moreover, inter-country comparisons make little sense, since they do not control for gravity-type effects. Some authors have looked directly at jobs reported to have fallen victim to offshoring, which seems much closer to the notion of offshoring, although there's bound to be some blur. Relating the number of such job losses to aggregate job losses over the relevant period (early 2000s) and in the countries considered (mostly the US), offshoring seems like a minor nuisance: In three of the four studies reported in OECD (2007), offshoring was involved in no more than 1 percent of the job losses.2 The EU has investigated job losses in a detailed analysis of 3,475 cases of industrial restructuring that took place in various member countries between 2003 and 2006. Offshoring to countries outside the EU has turned out as an ingredient of restructuring in 10 percent of the cases, and responsible for about 8 percent of the announced job losses. This suggests a somewhat bigger importance, but the authors still conclude that the scale of offshoring is smaller than might be expected. Interestingly, there is no evidence from this exercise that the phenomenon has gained importance during the years. But it is difficult to say what such numbers might tell us. May we conclude that all the discussion about offshoring is much ado about next to nothing? Or do these figures simply reflect an underutilized but large potential? A number of consultancies have ventured to estimate the jobs likely to be moved offshore. The numbers appear somewhat more impressive, between 1 and 9 percent of the estimated aggregate job losses, although still difficult to interpret without a clear benchmark. Taking a broader perspective, some others have even ventured to estimate the number of jobs that could potentially be moved offshore, mostly for the US.
2
A similar conclusion emerges from Belessiotis, Levin and Veugelers (2007).
20
Wilhelm Kohler
Expressing job losses as a percent of the relevant employment figure, offshoring now looks a more intimidating spectre, causing layoffs between 10 and 25 per cent of employment.3 If there is substantial disagreement about the quantitative significance of the phenomenon itself, this is aggravated by ambiguous messages about its effects. From an economic policy perspective, three dimensions seem important. One is welfare, the question of whether we may view offshoring as a phenomenon that enhances the gains from trade. The second is distribution, the main concern being its effect on domestic wages in the offshoring country. And finally, there is concern about displacement effects in industries where offshoring is observed. Economic theory suggests that the three dimensions are closely interrelated, but empirical studies have tended to focus on single aspects, mainly looking at either wage or employment effects. Unfortunately, however, the stories presented differ widely.4 In very broad terms, there are two reasons for this. One has to do with the ambiguity regarding measurement of offshoring itself; see the preceding paragraph. The second has to do with the lack of clear guidance from theory as to how, exactly, estimation equations should be specified. Both problems, particularly the second, are familiar from the literature on trade and wages. There are two principal modeling approaches to offshoring. One is to follow traditional trade theory, using general equilibrium models. These models may in turn be Ricardian or Heckscher-Ohlin in nature, taking a long-run view on comparative advantage, or they may be more short-run in nature, assuming factor specificity as in the Ricardo-Viner model.5 The strengths of these models is their ability to address welfare and distributional concerns, which also 3
4
5
See OECD (2007, p. 90). Blinder (2006) is particularly outspoken in identifying a large potential for offshoring, particularly in high-skilled labor-intensive tasks. A convenient survey is found in Belessiotis et al. (2006). See also Kohler (2007). This type of literature essentially goes back to Jones and Kierzkowski (1990). Important contributions are Feenstra and Hanson (1997), Jones (2000), Jones and Kierzkowski (2001a, 2001b), Deardorff (2001a, 2001b, 2005), Kohler (2003, 2004b), and Grossman and RossiHansberg (2006), which are all inspired by the Heckscher-Ohlin model. Kohler (2001, 2004b) and Bhagwati et al. (2004) take a Ricardo-Viner perspective.
Offshoring: Why Do Stories Differ?
21
seem to dominate the policy debate. This approach views offshoring as a result of some exogenous change in the cost of cross-border linking of production stages (cost of transportation and/or communication across distance and jurisdictions), which allows for a finer exploitation of cross country differences in the factor cost of performing different tasks in production. Such costs will in general also relate to the organizational mode of doing things, such as armslength transactions versus intra-firm transactions. However, in this first approach, these costs are typically treated as a `black box'. In contrast, the second approach to analyzing offshoring is to open this box, thereby also shedding light on the nature of transactional problems that are responsible for whether firms chose one mode of sourcing over another, in addition to determining where to source their inputs or tasks of production.6 Typically, however, models in this tradition are somewhat short on factor price effects and distribution, which is arguably a dominating concern of the policy debate. In this paper, the focus lies squarely on the first approach which places much emphasis on factor-cost considerations behind offshoring and on general equilibrium repercussions, but which takes a very simple view on the transaction-costs of offshoring, and which largely remains silent about the particular organizational form in which offshoring might take place. The results derived obtain independently of the specific organizational form. The model proposed is inspired by the Heckscher-Ohlin model of comparative advantage and trade. The purpose of the analysis is to shed light on the three different issues that have been addressed in the empirical literature, and the differing stories that exist where it has sometimes been difficult to interpret the results obtained, largely for lack of a lucid theoretical exposition of the effects in question. The first issue relates to the so-called productivity effect of offshoring. The second relates to the wage (or more generally factor price) effects of offshoring. And the third relates to employment (or more generally reallocation) effects of offshoring. The model will show that offshoring does indeed incorporate something like a productivity effect. It drives the gains from offshoring that many economists almost routinely emphasize when discussing the virtue of offshor6
Important contributions to this type of literature are McLaren (2000), Grossman and Helpman (2002, 2005), and Antras and Helpman (2004).
22
Wilhelm Kohler
ing, alluding to the general principle of gains from trade. However, the model also suggests that (and explains why) empirical studies should have a hard time identifying such productivity effects. As to the factor price and employment effects, the model will reveal that the a priori intuition that inspires much of the empirical work ex ante, and is usually invoked ex post in order to interpret the results obtained, is potentially misleading. The paper is structured as follows. Section II presents a simple two-sector model where production takes place in a continuum of production stages or tasks, each relying on high- and low-skilled labor, respectively, with a varying skill-intensity which is assumed to be given (Leontief-type technology). I describe the extensive margin of offshoring as an endogenous variable. Offshoring is driven by juxtaposing the factor-cost advantage of some foreign location where low-skilled labor is relatively cheap, and the extra cost of connecting production stages towards final goods production. I also introduce the concept of marginal skill intensity of an industry, marginal meaning at the extensive margin of offshoring, and the average skill intensity across the entire range of domestic tasks. Section III then turns to a general equilibrium analysis of a simple scenario which involves a reduction of the offshoring cost. Importantly, this is allowed to happen independently in both industries. I shall explore in some detail a key difference that arises between two different fundamental views of offshoring. One views offshoring as an input-related phenomenon, by definition applying to all industries at the same time, while the other views offshoring as a phenomenon which is idiosyncratic for each industry, and which always affects both types of labor within the industry. The former concept is used in the recent paper by Grossman and RossiHansberg (2006), whereas the latter has been used by Kohler (2003) and Kohler (2004b). It turns out that the wage effects of a reduction in the cost of offshoring are dramatically different for the two modes of offshoring. In section IV, I explore a scenario where domestic industries respond to a change in final goods prices, given exogenously from world markets. Does an industry contract or expand simultaneously at the intensive margin (meaning contraction or expansion of all existing domestic tasks) and at the extensive margin (meaning an expansion or reduction in the measure of tasks performed domestically, as opposed to offshore)? I introduce a concept of skill-intensity-difference between industries that allows us to tell whether an industry moves in the same direction on both
Offshoring: Why Do Stories Differ?
23
margins, or whether industries expand (contract) at the intensive margin, while contracting (expanding) at the extensive margin. Again, it will turn out that the two concepts of offshoring hold different messages. In section V, I shall draw some conclusions, with special emphasis on implications for empirical modeling. II. A Simple Model Any model of offshoring requires to be explicit about the way in which a certain production process may be decomposed, or fragmented, such that various parts may take place at two different locations (countries) featuring different factor costs. Factor costs may be different in two countries either because they have different technologies, or because they have different factor prices. I focus on factor price differences as determinants of factor cost advantages, although I do allow for technology differences to play a role as well. As regards factor cost, I assume that there is a foreign economy with given wage rates for high-skilled labor and lowskilled labor, respectively, denoted by wh and wl . I thus assume a two-factor setup with high-skilled and low-skilled labor as the only inputs. Labor endowments are considered as given exogenously for the larger part of my analysis.7 I also assume two final goods, where production is assumed to require a continuum of tasks, each requiring inputs of the two types of labor in a certain ratio, according to a Leontief-type production relationship. It will become evident that the model could be generalized to allow for factor substitution with relatively little effort, but since there are almost zero gains from doing so I stick to the Leontief-case. I use aih ( j ) and ail ( j ) to denote the fixed input coefficients per unit of task j in production of good i .8 The 7
8
The model is inspired by Dornbusch et al. (1980), Dixit and Grossman (1982), Feenstra and Hanson (1997), and the recent contribution by Grossman and Rossi-Hansberg (2006). See also Kohler (2004). The distinctive features of the model used here will be emphasized as I go along. Due to Grossman et al. (2006), it has now become standard to talk of tasks, where earlier literature has referred to stages of production, or fragments. Likewise, it has become costumary to use offshoring to describe foreign (as opposed to domestic) sourcing of tasks, leaving the precise mode of sourcing indetermined (in-house or outsourcing).
24
Wilhelm Kohler
continuous variable j [0,1] is used to index tasks, and i ^1, 2` is used to index the two goods or industries. Moreover, fi ( j ) denotes
the "amount" of task j that is required per unit of good i . Thus technology is also of the Leontief-type regarding the tasks, in addition to the types of labor used for each task. Essentially, f i ( j ) measures inverse task-productivities, or the importance of different tasks in production, across the continuum of tasks from j 0 to j 1 . By appropriate scaling of output-units, I assume a unitary 1
measure of tasks, i.e. ³ 0 f i ( j ) d j 1 . Notice that this does not imply a uniform distribution of inverse task-productivities (or importance) across j . Using wh and wl to denote the domestic economy's wage rate for high- and low-skilled labor, respectively, the factor-cost per unit of good i , barring any possibility of offshoring, is
ci ( wh , wl )
³
1 0
f i ( j ) > aih ( j ) wh ail ( j ) wl @ dj
(1)
Suppose now that there is a technology of linking tasks across distance. More specifically, if the home economy produces good i with some of the tasks performed in the other country, then the amount of task j that needs to be performed by foreign labor in order to secure availability of an equivalent to f i ( j ) domestic tasks, is ti ( j ) fi ( j ) , where ti ( j ) ! 1 is allowed to vary across tasks, differently across industries. This employs the notion of iceberg-cost to offshoring, capturing all costs involved in "gluing" tasks across locations.9 We do not dwell on details here, other than the fact that these costs vary across tasks. We now define J i ( wh , wl , j ) : > aih ( j ) wh ail ( j ) wl @ / > aih ( j ) wh ail ( j ) wl @ (2) as a measure of the factor cost of offshoring that derives from the factor intensity of tasks and the factor price difference between the domestic and the foreign economy. We know from multiple-goodversions of the Heckscher-Ohlin theory that ranking tasks in such a way that aih ( j ) / ail ( j ) is weakly increasing in j for both goods
9
See Baldwin (2006b). Jones and Kierzkowski (2005) call this the extra cost of service link activities.
Offshoring: Why Do Stories Differ?
25
implies monotonicity of J i ( , j ) in j . More specifically, for
wh / wl ! wh / wl the factor-cost-advantage J i ( , j ) is nonincreasing in j , and vice versa for wh / wl wh / wl .10 For wh / wl wh / wl , J i does not depend on j . We denote this level of J i by J i . If J i ! 1 , this indicates an across-the-board absolute cost advantage of the foreign economy. Without loss of generality we may normalize J i 1 . Cost-minimization requires that firms delocalize a task j , if
> aih ( j ) wh ail ( j ) wl @ ! > aih ( j ) wh ail ( j ) wl @ ti ( j ) ,
i.e.,
if
J i ( wh , wl , j ) / ti ( j ) ! 1 . There is a unique cut-off point separating tasks where cost-minimization dictates delocalization, or offshoring, and tasks to be performed domestically, provided that J i ( wh , wl , j ) / ti ( j ) is strongly monotonic in j . A convenient assumption often made is that it strongly decreases in j . The usual interpretation is that wh / wl ! wh / wl and the cost of offshoring,
t i ( j ) , rises in j , or at least that J i ( wh , wl , j ) / ti ( j ) inherits the factor-cost-based monotonicity from J i ( wh , wl , j ) . However, this may seem questionable. It is not the skill-intensity of a task per se that makes it more or less offshorable, but the degree to which it may be described in algorithmic terms, or the degree to which it requires face-to-face contact or complex communication. As argued by Blinder (2005, 2007a, 2007b) and others, it is by no means clear that ranking tasks in terms of "gluing-cost" replicates the ranking in terms of skill-intensity. Unless stated otherwise, in what follows, I assume that there is a rank-order of tasks such that J i ( wh , wl , j ) / ti ( j ) decreases monotonically in j , and that this rank order also implies that aih ( j ) / ail ( j ) is rising in j . I now denote the cost-minimizing cut-off value of j that satisfies J i ( wh , wl , j ) / ti ( j ) 1 (3)
10
See Jones (1956) and Dornbusch, Fischer and Samuelson (1980).
26
Wilhelm Kohler
by ji* ( wh , wl ) .11 In Kohler (2004b, 2007) I have called this the extensive margin of offshoring. It obviously depends on domestic factor prices. The cut-off level ji satisfies the first-order-condition for the unit-cost with offshoring, i.e.,
³ t ( k ) f ( k ) > a ( k ) w a ( k ) w @ dk ³ f ( k ) > a ( k ) w a ( k ) w @ dk
ci ( wh , wl , j ) :
j
0 i 1 j
with respect to
i
i
ih
ih
h
h
il
il
l
(4)
l
j . This first-order-condition simply requires
J i ( wh , wl , j ) / ti ( j ) 1 , and the second-order-condition is satisfied from the aforementioned monotonicity. It is instructive to consider the partial equilibrium comparative statics of ji* , holding factor prices constant. Thus, consider the case where ti ( j )
Ei zi ( j ) and
let there be an exogenous change 'E i 0 . For given domestic wages, this results in an increase in ji* which we could derive from the differential of the first-order-condition on ji* . Intuitively, the flatter J i ( wh , wl , j ) / ti ( j ) with respect to j , the stronger the reaction of the extensive margin of offshoring. In turn, a relatively flat schedule may arise from a flat factor intensity ranking, meaning a relatively homogeneous set of tasks and/or a flat ti ( j ) -schedule. We might call such an industry relatively sensitive with respect to offshoring. We may also explore offshoring effects of changes in factor prices. Any domestic wage change that increases costs at the marginal stage of production, aih ( ji* ) d wh ail ( ji* ) d wl ! 0 , makes the domestic economy uncompetitive at the margin and leads firms to delocalize further stages until the condition J i ( wh , wl , j ) / ti ( j ) 1 is reached. The opposite holds true for wage changes that satisfy aih ( ji* ) d wh ail ( ji* ) d wl ! 0 . Any factor price change that satisfies aih ( ji* ) d wh ail ( ji* ) d wl 0 leaves the extensive margin of offshoring unaffected. This property 11
This corresponds to the condition ( w / w ) / > E t ( I ) @ 1 in Grossman and Rossi-Hansberg (2006b).
Offshoring: Why Do Stories Differ?
27
will be important when we consider how offshoring behaves in the process of industrial restructuring. The schedule ji* ( wh , wl ) describes offshoring behavior of industry i , independently of the price of its output, because offshoring is purely a matter of cost-minimization. Given our assumptions, there is a lower bound of w l /w h for which ji 0 , and an upper bound for ji
1 . Industries may obviously differ in
their offshoring characteristics, and the schedule ji* ( wh , wl ) is a convenient way of describing these characteristic. For the sake of a simpler notation, I have abstained from indicating that ji* is also a function of foreign wage rates which we treat as parametric throughout this paper. It will prove convenient to introduce the distinction between the marginal and the average domestic skill intensity of an industry. The marginal skill-intensity is denoted by * * D i ( wh , wl ) : aih ª¬ ji ( wh , wl ) º¼ / ail ª¬ ji ( wh , wl ) º¼ , while the average
skill-intensity
is
denoted
by
D i ( wh , wl ) :
aih ( wh , wl ) / ail ( wh , wl ) , whereby 1
³
ais ( wh , wl ) :
ais ( j ) fi ( j )dj for s h, l
(5)
ji* ( wh ,wl )
The complementary definition of the average skill intensity of delocalized part of production is denoted by D i ( wh , wl ) : aih ( wh , wl ) / ail ( wh , wl ) , whereby
the
ji* ( wh , wl )
ais ( wh , wl ) :
³
ais ( j )ti ( j ) fi ( j )dj for s
h, l
(6)
0
Of course, the difference between the marginal and the average skill intensity vanishes as ji* approaches 1. Note that the coefficients ais are constant by assumption, while ais depends on domestic wages through the first-order condition on ji . With these definitions, we may now move from the partial industry perspective to general equilibrium. We look at a small open economy facing given world prices p1 and p2 for the two
28
Wilhelm Kohler
goods. With offshoring, the zero-profit-conditions are
pi > aih ( wh , wl ) wh ail ( wh , wl ) wl @ d aih ( wh , wl ) wh ail ( wh , wl ) wl for i 1, 2
(7)
with equality if there is any remaining home production in industry i. In line with the theory of effective protection, we may call the left-hand side of (7) the effective price for the domestic value added in industry i .12 Denoting this effective price by S i ( pi , wh , wl ), we arrive at conventional zero-profit-conditions of the following form:
S i ( pi , wh , wl ) d aih ( wh , wl ) wh ail ( wh , wl ) wl for i 1, 2
(8)
again with equality if industry i is at least partly viable domestically, meaning ji* ( wh , wl ) 1 . The key difference to the conventional zero-profit-condition is that instead of the price pi we now have an effective price which depends on domestic wage rates. It also depends on the technology of offshoring, ti ( j ) , which enters through the extensive margin ji , as well as through the function
ti ( j ) in the integral (6). It must be re-emphasized that the condition governing the extensive margin of fragmentation is influenced by the given wage rates in the foreign economy. In general equilibrium, these zeroprofit-conditions must be satisfied simultaneously for all industries. In addition, equilibrium requires full employment
a1s ( wh , wl ) y1 a2 s ( wh , wl ) y2
Ls for s
h, l
(9)
Equations (7) and (9) determine the two domestic wage rates as well as the gross-size of the two industries, y1 and y2 . The offshoring-schedules ji* ( wh , wl ) then determine the degree of delocalization in production of the two industries. III.
Gobalization: The Good of It
I now proceed to comparative static analysis. First, I consider the oft-quoted story of a more global village, which in the present 12
I have introduced this concept in Kohler (2003); see also Kohler (2004b and 2007).
Offshoring: Why Do Stories Differ?
29
modelling setup amounts to a reduction of ti ( j ) . As Grossman and Rossi-Hansberg (2006b) have emphasized, for this scenario to have interesting effects, one must assume offshoring to be present in the initial equilibrium. This generates infra-marginal effects. I assume that the cost of service link activities obey ti ( j ) Ei zi ( j ), and I assume Eˆi : 'E i / E i 0 . Hence, globalization increases the ease with which offshoring may take place across the board for all tasks involved in any one industry. I call it the "good of globalization", since it involves a true cost-saving, a reduction in real offshoringcost. In the next section, I will turn to a somewhat less benign form in which globalization presents itself, which is simply a fall in world-market prices of traded goods. However, I want to allow for Eˆ1 z Eˆ2 . This is a crucial point, much less innocuous than may appear at first sight. It has to do with a fundamental property of our offshoring technology, and a property which separates this model from Grossman and Rossi-Hansberg (2006). It warrants a brief digression. In this model, offshoring is an industry-specific phenomenon, relating to the idiosyncratic way in which the value added process of a certain industry may be sliced up, or fragmented, into different tasks. Each task, the smallest possible unit of the value added process, requires composite labor according to a Leontief-type technology. Accordingly, the gluing-cost for delocalized tasks are also an idiosyncratic element of an industry's technology. Any improvement in this glue (globalization scenario) is thus also an industry-specific element, but within the industry it affects all types of labor directly. Other industries may be affected indirectly through general equilibrium repercussions. By way of contrast, Grossman and Rossi-Hansberg (2006) define offshoring as an input-specific phenomenon, whereby the term input directly relates to a certain type of labor. If globalization allows easier gluing of input-specific tasks performed at different locations, then this directly affects all industries using this input, but it does not directly affect other tasks. Indeed, they are not affected indirectly, as we shall see below. These are two fundamentally different perceptions of what offshoring is all about. Both seem to have some merits empirically. The difference between them in terms of formal analysis is best illustrated in the present context if we make a further simplifying
30
Wilhelm Kohler
assumption. Let me thus assume that the input-coefficients for highskilled and low-skilled labor are the same across tasks, i.e., ais ( j ) ais for s h, l and i 1, 2 . This assumption, also made by Grossman and Rossi-Hansberg (2006), is enormously helpful analytically, since it allows us to substitute for foreign factor-cost, using the first-order-condition on the extensive margin ji* , which now reads as
aih wh ail wl (10) Ei zi ( ji* ) Notice that we have now replaced ti ( j ) Ei zi ( j ) . I still denote aih wh ail wl
the endogenous margin of offshoring by the schedule ji* ( wh , wl ), which depends parametrically on foreign wages wh and wl . Replacing accordingly in (5) and (6), we obtain the following zeroprofit-conditions: pi aih wh ail wl Si ª¬ ji* wh , wl º¼ for i 1,2 (11) where
1 Si ª¬ j wh , wl º¼ : zi ( ji* ) * i
ji* ( wh , wl )
³ 0
1
zi ( j ) f i ( j )dj
³
ji* ( wh , wl
fi ( j )dj (12) )
The schedule Si ª¬ j wh , wl º¼ captures how changes in the technology of offshoring affect the total factor cost of industry i through a change in the cost-minimizing extensive margin ji* . A number of things are worth emphasizing about this.13 Due to the second-order-condition on ji* , we have Si 1 , which reflects the simple fact that offshoring leads to net savings on factor cost. Moreover, note that the shift parameters E i have disappeared through substitution, which may seem counter-intuitive. However, they still play a role through determining the equilibrium margins of offshoring in the two industries, j1* and j2* . It is obvious that * i
Sic 0 , and I use Z i : Sic ji / Si 0 to denote the elasticity of the 13
This discussion relies on Grossman and Rossi-Hansberg (2006). For a comparison of results see below.
Offshoring: Why Do Stories Differ?
31
offshoring-savings-factor Si with respect to ji . And finally, for ji* 0 we have Si 1 , due to the assumed unitary measure of
fi ( j ) . According to (10), for constant foreign wage rates, the comparative statics of the extensive margin satisfies
ˆji*
1
]i
T
ih
wˆ h Til wˆ l Eˆi
(13)
where ] i : zi' ji / zi ! 0 denotes the elasticity of the gluing-cost function zi ( j ) which now drives all movements at the extensive margin, due to the assumption of otherwise homogeneous tasks. Note that this elasticity is evaluated at the initial equilibrium value of ji* . I use Tis to denote the cost-share of factor s , as usual. Assuming a small country facing constant final goods prices p , the zero profit conditions (11) imply
Tih wˆ h Til wˆ l Zi ˆji*
0
(14)
Combining these two equations, our globalization scenario now emerges as
Zi
] i Zi
Eˆi
Tih wˆ h Til wˆ l
for i 1, 2
(15)
At first sight, the comparative statics may seem ambiguous, since ] i ! 0 and Zi 0 are of opposite signs. However, these two elasticities are intimately related to each other, and it is relatively straightforward from plain intuition that ] i ! Zi . The elasticity
] i measures the extent to which expanding the scope of offshoring increases the marginal offshoring-cost zi ( j ) . The elasticity Zi measures the extent to which doing so increases the entire factor cost savings from offshoring, relative to all-home-production, measured through the offshoring-savings-factor Si 1 . Both elasticities are evaluated for equilibrium values where the firstorder-condition on ji* is satisfied. This condition requires that the marginal effects would just offset each other, but since Si includes the infra-marginal savings effects which are smaller (due to the
32
Wilhelm Kohler
assumption that ] i' ( j ) ! 0 ), the marginal effect measured by ] i must dominate the effect on Si in absolute value: ] i ! Zi , whence Zi / ] i Zi 0 . It should be noticed that the left hand side of (15) is the percentage increase of the effective price, as defined in (8), that is brought about by a Eˆi -percent reduction in the offshoring cost. Thus, the productivity effects may equivalently be expressed as an increase in the effective price.14 Equation (15) is intuitive, both directly and also from the notion of an effective price change. If firms in a certain industry already engage in offshoring, i.e., if ji* ! 0 , any across the board reduction in gluing-costs through Eˆi 0 acts just like an increase in the price of the good, given that Zi / ] i Zi 0 , as demonstrated before.15 An alternative way to describe the same effect is to say that offshoring makes domestic factors used in an industry uniformly more productive. The strength of the effect depends on how far the industry has already gone in terms of the scope of offshoring. This is because of the inframarginal effect of making all those tasks cheaper to obtain that have already been sourced abroad, prior to the fall in E i . If fi ( j ) is uniform, this effect is measured through the margin ji* alone. If fi ( j ) varies across j , meaning that some tasks are inherently more important for production than others, then this effect is the larger, the more important the tasks that have been delocalized to start with. Moreover, the strength of the effect depends on the steepness of the zi ( j ) -schedule. Taken together, the strength of this equivalent-price-increase for a given size of the shock is given through the term Zi / ] i Zi 0 . As regards the general equilibrium effects, we may now invoke Stolper-Samuelson logic to pin down the wage effects of offshoring, provided that the economy is and remains diversified. 14
I have emphasized this perspective already in Kohler (2003, 2004), but without offering an explicit solution for Sˆi .
15
See Kohler (2003) where I have first pointed out that offshoring may be seen in this way.
Offshoring: Why Do Stories Differ?
33
The diversification condition matters because any change in ji* also affects the full employment conditions, since less domestic labor is now used to generate outputs y1 and y2 . Under our assumption of constant input coefficients across all tasks, both types of labor are affected symmetrically. Offshoring just acts like a sector-biased technological change, which may equivalently be treated as an effective change in the supply of both types of labor. In our case, this happens in a disproportional way, but as long as this does not push the economy outside its cone of diversification, the wage effects are determined in the familiar way from the zero-profitconditions alone. With a view on the fundamental distinction between two different concepts of offshoring introduced above, it should now be instructive to compare this expression with the corresponding result in Grossman and Rossi-Hansberg (2006). To have a meaningful comparison, we need to look at the more general case where they allow for trade in tasks for both low- and high-skilled labor. For this case, their zero-profit-conditions read as (in terms of our notation)
aih wh Sh ª¬ jh* ( wh ) º¼ ail wl Sl ª¬ jl* ( wl ) º¼ for i 1, 2 (16)
pi
where S s is defined by analogy to (12), the only difference being that they assume f s ( j ) to be uniform with measure 1, now relating to input s
^h, l` ,
and not to industry i
The comparative statics on
ˆjs*
wˆ Eˆ / ] , where ] s
s
s
s
j
s
^1, 2` ,
as in our model.
in their case simply require
is defined by analogy to ] i above.
For constant goods prices pi the two equations (16) uniquely determine equilibrium domestic factor costs for each of the two factors,
high-
and
low-skilled
labor,
wh Sh ª¬ jh* ( wh ) º¼
and
wl Sl ª¬ jl* ( wl ) º¼ , respectively. Hence, comparative statics must obey wˆ s
Sˆs , independently for both s
h and s l . Offshoring of the two types of tasks takes place with no connection whatsoever between these two phenomena.16 This is a stark separation property 16
See section 4 of Grossman and Rossi-Hansberg (2006b) for more details.
34
Wilhelm Kohler
for input-specific offshoring which, obviously, does not hold for industry-specific offshoring. What are the conclusions to be drawn from this analysis? A first point to be made in view of the counter-intuitive results that I have emphasized in Kohler (2003) and Kohler (2007) is that these are ruled out here because of the continuity imposed on offshoring through assuming a continuum of tasks. The above analysis looks at small changes where all pathologies behind counter-intuitive results are ruled out. However, with a view on the empirical literature it must still be emphasized that some of the direct intuition often invoked is misleading. As has often been pointed out by trade economists in the debate about trade and wages, in open economies wages are determined by prices through the zero-profit-conditions, more than by quantities - say quantities traded. In the present context a similar point obtains for offshoring. More specifically, we may have an industry where in a given period (sample) a lot of action takes place in terms of going offshore, and we may be confident that it is all a story about Eˆi 0 . Yet, if this happens at a young age of offshoring, with ji* close to zero at the outset, then we should not be surprised to find small effects on wages. In other words, when we estimate the role of offshoring through an empirical implementation of zero-profit-conditions like (15) above, we should also attempt to observe, by whatever proxy might be available, the terms Zi / ] i Zi across industries. And it does matter from the above whether offshoring is of one or the other type, tasks that are idiosyncratic for industries, or tasks that are idiosyncratic to certain types of labor. For the latter type of offshoring, we do have a somewhat counter-intuitive result that offshoring of tasks that strictly require low-skilled labor raises the domestic wage for low-skilled labor. The same applies, independently, for offshoring of strictly high-skill-labor-tasks. As with other results that appear counter-intuitive at first sight, closer inspection reveals Stolper-Samuelson logic at work. If the effective cost of domestic low-skilled labor falls, because of cheaper availability of low-skilled-labor-tasks sourced abroad, this benefits the low-skilled-labor-intensive industry more than the other industry, and the only way to restore zero-profits is through an offsetting rise in the domestic wage for low-skilled labor. Again,
Offshoring: Why Do Stories Differ?
35
the same logic applies independently for high-skilled-labor-tasks. And in either case the effect would be much smaller at early stages of offshoring than for later stages where ji* is already large. A further point relates to welfare and real wages. With strictly input-specific offshoring, we would always observe a rise in real wages for both types of labor. Offshoring is a Pareto-improving event. The same does not hold true for industry-specific offshoring where the Stolper-Samuelson logic applies in a more familiar way. For this type of offshoring to be a Pareto-improvement, it would have to be sufficiently symmetric in terms of EˆiZi / ] i Zi being not too different in absolute magnitude across industries. And finally, as regards the productivity effect that has also been the object of empirical analysis,17 in this model all productivityequivalent gains are absorbed by factor price changes. It is thus difficult to imagine how productivity gains would show up in an empirical analysis. Of course, there is much more to data than this model can identify. In particular, productivity effects at the industry level may be due to firm heterogeneity as in Melitz (2003), which is ruled out here altogether. However, there is a certain similarity between the new literature on heterogeneity-driven productivity effects that work through firm selection. As I have suggested in Kohler (2007), we may re-interpret our ti ( j ) s schedule as incorporating a varying degree of Ricardian comparative advantage, driven by productivity-differences, across different tasks, in addition to the offshoring-cost emphasized above. In this case, enhanced globalization also elicits selection of, and enhanced concentration on, tasks where domestic firms have a larger degree of comparative advantage. Indeed, we may even view our continuum of tasks as also involving a continuum of different firms. The question then remains how to define productivity in such a way that we may detect a productivity effect of offshoring, not in addition to factor price effects, but as a different, but equivalent, way of observing the same phenomenon - realizing that the factor price effects are essentially mandated from the productivity effects of offshoring.
17
See for instance Mann (2003), Mann and Kirkegaard (2006), Amiti and Wei (2005), and Görg and Girma (2004).
36
Wilhelm Kohler
IV.
Globalization: A Less Benign Interpretation
In the preceding section, the globalization scenario was essentially good news since it has involved easier access to cheap things from abroad. In this section, I want to take a somewhat less benign perspective on globalization. Thus, suppose that due to enhanced world supply of good 1, its price falls on world markets. Domestic firms in industry 1 will perceive this as an increase in competitive pressure, and they will have to adjust. For the economy as a whole, this requires an adjustment in wage rates, as well as factor reallocation across industries which may be costly, and may involve temporary unemployment. From a general equilibrium perspective, in a two-goods-model it only makes sense to look at the relative price. Without loss of generality, I shall therefore look at a fall in the relative price of good 1 by normalizing p2 1 . I do not want to take a stance as to whether this constitutes a terms-of-trade improvement or a terms-of-trade deterioration, although one can easily portray the scenario in that way by assuming an appropriate demand and trade pattern. However, my story is not a terms-oftrade story, but a story of structural adjustment of a country's production, dictated by some exogenous change in world market prices. Of course, the story might just as well be told as one of a price increase, in which case it might look more benign. But globalization here is still less benign than in the previous section, simply because it now is a two-sided coin, as each relative price change is, whereas before it was one-sided in that there was an improvement in technology. The issue that I want to address is how long-run adjustment of domestic production evolves in the face of secular changes of comparative advantage, brought about by an exogenous change in relative prices of traded goods. Of course, adjustments in quantities and prices are not independent. Both sides adjust simultaneously and in interdependence. Indeed, if wages are perfectly flexible, then the wage adjustment might seem more important for policy than the associated adjustment in quantities in terms of reallocation and industrial restructuring. But there are important reasons why the quantity side is important to look at in its own right. First, if the reallocation effects associated with adjustment under flexible prices are large, then the unemployment effect that would arise in case of wage rigidities is also large. Looking at the quantity side of
Offshoring: Why Do Stories Differ?
37
adjustment is, thus, a first and rough way to gauge the potential unemployment problem that may arise from a certain scenario if wages are rigid. The second reason has to do with adjustment costs, which are often assumed away. Short of explicitly modeling such costs, the magnitude of adjustment in quantities that arise from a given scenario may give a first and rough indication of the likely costs of adjustments. An issue that is not usually addressed, but which may be important, is whether the long-run decline of a certain industry in terms of employment goes hand in hand with a gradual reduction also in production depth. One might be inclined to assume that it does, but the subsequent analysis reveals that the opposite is also possible, whereby an ever smaller number of people employed in an industry carries out an ever larger set of tasks, covering an ever larger range of production stages. In a similar vein, if a country acquires comparative advantage in some industry, will this necessarily happen in the form of a gradual increase not only in employment, but also in the number of production stages carried out domestically? Or might the industry increase in size (employment) through specialization on an ever smaller subset of production stages? Arguably, adjustment costs may be quite different in the two forms of restructuring. I shall not, however, model such costs explicitly in this paper. With Leontief-type technology, goods price changes normally do not entail any reallocation at all. Under flexible prices, all adjustment takes place in factor price space, until - in the extreme case - one of the two factors has a zero shadow value and production adjusts in a discrete way towards complete specialization. To allow for a more interesting story of quantity adjustment, we need to return to the more general model where there is a systematic variation of skill-intensity across the continuum of tasks in both industries considered. In this case, offshoring makes the skill-intensity variable. But even if tasks are homogeneous in terms of skill-intensity, the conclusion from the above analysis would be that offshoring introduces an element of continuity into adjustment through the rising cost-schedule t i ( j ) . Moreover, the previous analysis seems to suggest that an exogenous price increase (reduction) should have the effect of a reduction (increase) in offshoring. This follows from the fact that offshoring is the equivalent, in terms of the zero-profit-conditions, of a price increase. Hence, a price reduction may partly be undone through
38
Wilhelm Kohler
offshoring. However, in this section I want to show a special feature of adjustment that arises in the more appealing case where tasks are heterogeneous in terms of their skill-intensity. It will prove useful for the subsequent analysis to introduce the following definitions, relying on the notions of average and marginal skill-intensity introduced in (5) above. I shall call industry 1 strongly more skill-intensive than industry 2, if it features both, a higher average and a higher marginal skill intensity than industry 2, for given external margins of offshoring, j1* and j2* . I call industry 1 weakly more skill intensive, if at the given levels of ji its average skill intensity is higher, but its marginal skill intensity is lower than in industry 2's average skill-intensity. And finally, industry 1 is called globally more skill intensive than industry 2, if its marginal skill intensities over the entire range of ji [0,1] span a skillintensity-cone which is disjoint from, and more skill-intensive, than the corresponding cone for industry 2. It turns out that with this definition, important qualitative results may be obtained without having to work out a full closed-form-solution of the comparative statics. Moreover, the difference between our two forms of offshoring are also relatively easy to establish. I first look at the case of industry-specific offshoring with heterogeneous tasks, followed by a brief treatment of the same scenario for the alternative concept of offshoring proposed by Grossman and RossiHansberg (2006). It is important at the outset to note that our scenario is an exogenous reduction in the relative price of a good which is relatively skill-intensive in the average sense. In the conventional scenario we would expect a decline in employment of both types of labor in this industry, and a reallocation towards industry 2. This happens through the equilibrium wage adjustments which make high-skilled labor relatively cheaper, thus making both industries more skill-intensive. And full employment then requires a contraction of the more skill-intensive of the two industries. As a first step, let us thus look at notional wage adjustments that would maintain the zero-profit conditions in the face of dp1 0 without any change in the margin of offshoring. Assuming diversification, using (7) this type of wage adjustment is described by dp1 a1h ( wh , wl )dwh a1l ( wh , wl )dwl 0 (17a)
Offshoring: Why Do Stories Differ?
39
0 a2 h ( wh , wl )dwh a2l ( wh , wl )dwl
(17b) where the second equation explicitly states that the price of good 2 has been normalized to p2 1 . Note that by construction of my argument, the factor cost for delocalized tasks on the left-hand side of (7) remains unchanged. We might call it the incipient wage adjustment. Now suppose that industry 1 is weakly more skill-intensive than industry 2 at the initial levels of offshoring ji* and domestic wage rates,
wh
and
wl , respectively. This means that
a1h () / a1l () ! a2 h () / a2l () ! a1h ª¬ ji* ( wh , wl ) ¼º / a1l ¬ª ji* ( wh , wl ) ¼º , where the average skill-intensities are defined as in (5) above, and where a1s ª¬ ji* ( wh , wl ) º¼ denotes the s -type labor input-coefficient at the initial margin ji* , with s h, l . Remember that ji* satisfies the first-order-condition (10) for cost-minimization through localization of tasks. Obviously, under these skill-intensity assumptions, the wage changes from (17) feature a fall in wh and a rise in wl , such that minimum-cost of industry 2 remains constant. Since the marginal task of industry 1 features a lower skill-intensity than industry 2 does on the average of its tasks, this implies that the factor cost of that task must rise. The reason is that a lower skill-intensity makes it less well positioned to gain from a relative reduction of the highskilled wage rate. From (10) it then follows that these notional wage adjustments are no equilibrium reactions, since they violate the first-order-condition. More specifically, the rise in w l and fall in wh are such that the domestic economy loses its competitive edge on the marginal task ji of the skill-intensive industry 1, and cost-minimization requires that further tasks are delocalized offshore. The same holds true for the less skill-intensive industry 2, since by definition each industry's average skill-intensity is larger than its marginal skill-intensity. In other words, equilibrium adjustment in this case requires enhanced offshoring in both industries. Both industries thus become more skill-intensive in the process of adjustment, although for small changes industry 1 remains the more skill-intensive of the two on the average of all tasks performed domestically. The equilibrium wage adjustment is in line with the Stolper-
40
Wilhelm Kohler
Samuelson logic, although different from that described in equations (17) above.18 The opposite holds true for industry 1, if it is strongly more skill-intensive than industry 2, and a fortiori if it is globally more skill-intensive. In this case, we observe a partial reversal of offshoring in industry 1, since notional the wage adjustment according to (17) now implies that industry 1 actually gains a competitive edge on the marginal task. The opposite still holds true for industry 2 (as in the previous case), by definition of the average skill-intensity in (5). Hence, equilibrium adjustment must now involve less offshoring in industry 1 and more offshoring in industry 2. As a result, industry 1 becomes less skill-intensive, while industry 2 becomes more skill-intensive. The factor-intensity effect of offshoring thus reduces the skill-intensity difference between the two industries. This, in turn, implies that the Stolper-Samuelson logic, although still in force, implies a somewhat mitigated adjustment, compared to a case where offshoring is ruled out. As regards overall resource allocation, we lose the Rybczynski-type logic which normally implies contraction of industry 1 and expansion of industry 2. The reason is that the factor intensities of the two industries now move in opposite directions. To draw a somewhat more general conclusion, it is important to realize the specific nature of the scenario considered. In our model, industry 1 might be called a comparative advantage industry, because it is relatively skill-intensive on the average of tasks and the home economy has relatively cheap high-skilled labor. This industry faces increased competitive pressure from a fall in the price of good 1. Somewhat paradoxically, the conventional type of adjustment obtains, if the industry is the less skill-intensive at the margin of offshoring than the other industry is on average. The intuition is that in this case the incipient rise in wl and fall in wh harms the industry also at its offshoring margin. If the industry is skill-intensive also at this margin, then the incipient wage effects lead to opposite movements in the skill-intensities of the two industries that mitigate the Stolper-Samuelson logic and potentially negate the Rybczynski-logic for factor reallocation. 18
In Kohler (2004b), I present a similar analysis using a dual version of a somewhat simpler model where offshoring takes place only in one of the two industries.
Offshoring: Why Do Stories Differ?
41
Suppose that we have an exogenous relative price increase of industry 1, rather than a price reduction. Then the wage effects described by (17), with d p1 ! 0 , now go the other way, with a rise in wh and a fall in wl . Reiterating the logic of the preceding argument, we would conclude that if industry 1 is weakly more skillintensive, its reaction would involve a partial reversal of offshoring. The reason is that it is relatively less skill-intensive in terms of the task at the relevant margin of offshoring, hence it is less affected by the changes d wh ! 0 and d wl 0 from (17). The same holds true, by definition of average skill intensities as in (5), for industry 2. From Rybczynski-type logic, industry 1 expands, while industry 2 contracts. If industry 1 is strongly more skill-intensive, then these same wage adjustments would imply that industry 1 loses its competitive edge at the initial margin of offshoring, and would therefore react by delocalizing more tasks offshore, while the opposite holds true for industry 2. Thus, adjustment implies an increase in factorintensity differences. The result is that the Stolper-Samuelson logic gets reinforced, and the Rybczynski-type logic is lost, potentially at least. The conclusion to be drawn from this analysis for empirical work is that it makes relatively little sense to search for an unambiguous relationship between changes in the extent of an industry's level of offshoring, and changes in employment or more generally its level of activity. Everything depends on the type of exogenous shock that drives the data. Offshoring and employment, together with the wage effects, are all jointly endogenous, and ignoring this endogeneity in the estimation process, without bothering about the type of exogenous shock that generates the variation observed in the data, leads to potentially meaningless results. The preceding analysis indicates a rich pattern of possible industrial restructuring in a country that faces a change in world prices for traded final goods, and whose firms engage in offshoring to countries where low-skilled labor is relatively cheap. The underlying assumption was that offshoring is an industry-specific phenomenon involving heterogeneous tasks. A somewhat simpler pattern of restructuring obtains, if offshoring is specific to inputs as proposed by Grossman and Rossi-Hansberg (2006). Without going into details, I want to show in what follows why things are different and simpler in this alternative world. Changes in relative wages within the domestic economy, a key channel of the effects consid-
42
Wilhelm Kohler
ered in this section, are entirely irrelevant for offshoring. Industrial restructuring caused by price changes does involve changes in offshoring, however, also in the GR-model. From (16), the domestic factor costs for tasks, wh S h ª¬ jh ( wh ) º¼ and wl Sl ª¬ jl* ( wl ) º¼ responds in the familiar way to goods price changes. With given foreign wage rates and given offshoring margins ji* , this disturbs the offshoring condition. By necessity, it does so in both sectors, and the offshoring margins of the two industries will always move in the same direction. More specifically, from (16), the factor cost changes following a change in p1 must satisfy the familiar magnification relationship of the Stolper-Samuelson theorem. I.e., under the assumption that industry 1 is relatively skill intensive, pˆ1 0 must be associated with a fall in wh S h ª¬ jh* ( wh ) º¼ , and a rise in wl Sl ª¬ jl* ( wl ) ¼º . Now, since the extensive margins js* are increasing in the respective wage rates, and S sc 0 , and since we now assume a constant offshoringcost and a constant foreign wage rate, a fall in wh will be associated with a rise in S h , i.e., a partial reversal of offshoring. This is intuitive, since a lower domestic wage for high-skilled labor makes the domestic economy competitive at the initial margin of offshoring. Whether or not this is an equilibrium adjustment depends on ˆ h Zh ˆjh* 0 , with wˆ h 0 . From the first whether it is true that w order condition of offshoring, which in this case states that
wh
wh / ª¬ E h zh ( jh* ) º¼ , we have ˆjh*
is satisfied, if
wˆ h / ] h . Hence the condition
wˆ h (1 Zh / ] h ) 0 , which in turn implies
Zh / ] h ! 1 , which is equivalent to ] h ! Zh , or ] h ! Zh . Now, we know that ] h ! 0 and Zh 0, and in section III above we have also shown that the equivalent of condition ] h ! Zh is satisfied for the industry-specific notion of offshoring. By analogy, it is also satisfied in the present context. From all of this it then follows that, ˆ h 0 with for an equilibrium adjustment to pˆ1 0 , we do have w
Sˆh ! 0 , and conversely for low-skilled labor where wˆ l ! 0 and
Offshoring: Why Do Stories Differ?
43
Sˆl 0 . We may thus conclude that the Stolper-Samuelson theorem remains valid in qualitative terms for the Grossman-Rossi-Hansberg world of offshoring.19 For the reallocation effects it is now important to realize that the full employment conditions are given by a1h Sh ª¬ jh* ( wh ) º¼ y1 a2 h Sh ª¬ jh* ( wh ) º¼ y2 Lh (18a)
a1l Sl ª¬ jl* ( wl ) º¼ y1 a2l Sl ª¬ jl* ( wl ) º¼ y2
Ll
(18b)
From the previous paragraph we know that S h is rising, while
Sl is falling, in identical ways for both industries. Denoting the familiar employment shares by Ois : ais S s ª¬ js ( ws ) º¼ yi / Ls , we may write the relative changes as
O1h yˆ1 (1 O1h ) yˆ 2 O1l yˆ1 (1 O1l ) yˆ 2
Sˆh 0 Sˆ ! 0
(19a)
(19b) From this, we may directly conclude that output levels respond as in the Rybczynski-type magnification effect: yˆ1 Sˆh 0 l
Sˆl yˆ 2 . This is as expected. However, it relates to final output levels, not to employment of labor in the two industries. As emphasized above, overall employment in industry 1 falls, as perhaps expected, if both industries become more skill-intensive in the process of adjustment, with the skill-intensity measured as a S ª j * ( w ) º / a S ª j * ( w ) º . Since Sˆ 0 Sˆ , i.e., since ih
h
¬
h
h
¼
il
l
¬
l
l
¼
h
l
both industries conduct more offshoring of low-skilled tasks and a partial offshoring-reversal of high-skilled tasks, this condition is unambiguously satisfied, and the employment reaction is in line with the traditional model and plain intuition. More competitive pressure on the final output side of industry 1 leads to a contraction, not only of gross output, but also of domestic employment of both factors in that industry. The reverse happens in industry 2. Notice that we are experiencing a certain degree of gradualism in resource reallocation where without offshoring we would have no reaction up to a point, with a subsequent discrete jump to complete specialization. 19
See also Baldwin and Robert-Nicaud (2007).
44
Wilhelm Kohler
V. Conclusion
Stories of offshoring differ for two non-trivial reasons. First, the underlying view of the offshoring phenomenon may differ, and secondly, the scenario looked at in theoretical models, or the exogenous changes behind the data used in empirical work, may differ. If these differences are made explicit, then the different stories should not cause confusion. However, in empirical work at least, this is often not the case. And theoretical models usually subscribe, more or less arbitrarily, to a specific, single notion of offshoring. Moreover, they mostly look at scenarios where offshoring arises (sometimes from a case without offshoring to start with) due to an exogenous reduction of offshoring-costs. Offshoring as an ingredient to adjustment of other types of shocks has received little attention. The empirical literature very often investigates the effects of offshoring on variables like wages and employment which are jointly endogenous with offshoring itself. If two variables are jointly endogenous, their co-movement depends on the type of exogenous shock that drives this movement. Unless the researcher is explicit about this, results are difficult to interpret, and one should not be too surprised that the stories reported differ. In this paper, I have identified two different types of offshoring that are both amenable to general equilibrium analysis featuring several types of labor and may, thus, be compared in a rigorous way for the same type of scenario, focusing on distribution and factor allocation. Moreover, I have looked at two different scenarios to exemplify that the co-movement of offshoring on the one hand, and wages and employment on the other, heavily depends on the type of underlying shock. Offshoring may be related to tasks that are specific to certain types of labor. All types of labor may be subject to offshoring, but the tasks of low-skilled labor may be delocalized independently from the tasks performed by high-skilled labor. This is the new paradigm of trade proposed recently by Grossman and RossiHansberg (2006). Alternatively, offshoring may be seen as an industry-specific phenomenon where the smallest unit of production that may be delocalized still involves bundles of many (or all) types of labor. It is then impossible to delocalize inputs of low-skilled labor independently of high-skilled labor. This type of offshoring has been looked at in earlier literature, but the two types have not been analyzed and compared in a unified framework.
Offshoring: Why Do Stories Differ?
45
I have analyzed these two phenomena in a unified way, relying on Grossman and Rossi-Hansberg (2006), focusing on factor cost savings of delocalization. But in addition to their scenario of a lower offshoring-cost, I have also looked at a case where offshoring is part of an industrial adjustment caused by a change in prices for tradable final goods, given exogenously from world markets. Among the results obtained, offshoring of low-skilled tasks always raises the wage for low-skilled labor, as emphasized by Grossman and Rossi-Hansberg (2006), independently on factor intensity relationship between industries concerned. However, for the other type of offshoring factor intensity rankings do play a role, whereby a key distinction arises between the marginal and the average skill-intensity of an industry. More interesting differences arise if one looks at the employment and factor price effects of a change in final goods prices. More specifically, if offshoring involves heterogeneous tasks involving bundles of both types of labor, then contraction or expansion of an industry caused by such price changes may take place in the same direction at both the extensive and intensive margin of adjustment. But, depending on the skill intensity difference between the industries concerned, contraction of total employment in an industry may go hand in hand with expansion at the extensive margin, meaning a partial reversal of offshoring. Indeed, adjustment no longer needs to be in line with the familiar Rybczynski-type magnification effect. However, adjustment takes more mainstream forms if offshoring is of the other type where tasks relate to single types of labor. For empirical work, my analysis suggests that the specification of estimation equations for both, wage and employment effects, and the specific estimation technique chosen, should be sensitive with respect to the fact that offshoring, as well as wages and employment, are all jointly endogenous. Moreover, the empirical approach should allow for a distinction between the two types of offshoring. For the industry-specific form of offshoring, the specific ways in which industry-characteristics determine the wage and employment effects identified in the present analysis should be, and can be, taken into account in the empirical specification.
46
Wilhelm Kohler
References
Antràs, P., Helpman, E. (2004), Global Sourcing, Journal of Political Economy 112, 552-580. Amiti, M., Wei, S.-J. (2005a), Fear of Service Outsourcing: Is It Justified?, Economic Policy 20, 307-347. Amiti, M., Wei, S.-J. (2005b), Service Offshoring, Productivity, and Employment: Evidence from the United States, IMF Working Paper 5/238. Baldwin, R. (2006b), Globalization: the great unbundling(s), prepared for the Finnish Prime Minister's Office, Economic Council of Finland, (in context of EU Presidency 2006). Baldwin, R., Nicoud, F.R. (2008), Trade and Growth with Heterogenous Firms, Journal of International Economics 74(1), 21-34 (January 2008). Belessiotis, T., Levin, M., Veugelers, R. (2007), EU competitiveness and industrial location, European Commission, Bureau of European Policy Advisers, Luxembourg: European Communities. Bhagwati, J., Panagariya, A., Sirinivasan, T. N. (2004), The Muddles over Outsourcing, Journal of International Economics 14, 209-221. Blinder, A.S. (2005), Fear of Offshoring, Princeton University, CEPS Working Paper No. 119, December 2005. Blinder, A.S. (2006), Offshoring: The Next Industrial Revolution, Foreign Affairs 85 (2). Blinder, A.S. (2007a), How Many U.S. Jobs Might Be Offshorable?, Princeton Univerisity, CEPS working paper no. 142, March 2007. Blinder, A.S. (2007b), Offshoring: Big Deal or Business as Usual?, Princeton University, CEPS working paper no. 149, June 2007. Dixit A. K., Grossman, G. M. (1982), Trade and protection with multistage production, Review of Economic Studies 49, 583-94.
Offshoring: Why Do Stories Differ?
47
Deardorff, A. V. (2001a), Fragmentation in Simple Trade Models, North American Journal of Economics and Finance 12, 121137. Deardorff, A.V. (2001b), Fragmentation across cones, in: S. W. Arndt and H. Kierzkowski (eds.), Fragmentation: New Production Patterns in the World Economy, Oxford, Oxford University Press, 35-51. Deardorff, A.V. (2005), A trade theorist's take on skilled-labor outsourcing, International Review of Economics and Finance 14, 259-271. Dornbusch, R., Fischer S., Samuelson P.A. (1980), HeckscherOhlin Trade Theory with a Continium of Goods, Quarterly Journal of Economics 95, 203-224. Feenstra, R. C., Hanson, G.R. (1997), Foreign direct investment and relative wages: evidence from Mexico's Maquiladoras, Journal of International Economics 42, 371-93. Görg, H., Girma, S. (2004), Outsourcing, Foreign Ownership and Productivity: Evidence from UK Establishment Level Data, Review of International Economics 12, 817-832. Grossman, G. M., Helpman, E. (2002), Integration versus outsourcing in industry equilibrium, Quarterly Journal Economics 1117, 85-120. Grossman, G.M., Helpman, E. (2005), Outsourcing in a Global Economy, Review of Economic Studies 72, 135-159. Grossman, G. M., Rossi-Hansberg, E. (2006a), The Rise of Offshoring: It's not Wine for Cloth Anymore, Presented at the Federal Reserve Bank of Kansas City Economic Symposium The New Economic Geography: Effects and Policy Implications, Jackson Hole, Wyoming, August 24-26, 2006. Grossman, G. M., Rossi-Hansberg, E. (2006b), Trading Tasks: A Simple Theory of Offshoring, Mimeo, Princeton, N.J.: Princeton University. Jones, R. W. (1956), Factor proportions and the Heckscher-Ohlin theorem, Review of Economic Studies 24, 1-10. Jones, R.W. (2000), Globalization and the Theory of Input Trade, Cambridge, Mass., MIT Press.
48
Wilhelm Kohler
Jones, R.W., Kierzkowski, H. (1990), The role of services in production and international trade: a theoretical framework, in: R. W. Jones and A.O. Krueger (eds.), The Political Economy of International Trade, Oxford, Basil Blackwell. Jones, R. W., Kierzkowski, H. (2001a), A Framework for Fragmentation, in: S.W. Arndt and H. Kierzkowski (eds.), Fragmentation: New Production Patterns in the World Economy, Oxford, Oxford University Press. Jones, R.W., Kierkowski, H. (2001b). Globalization and the consequences of international fragmentation, in: Calvo, G., Obstfield, M., Dornbusch, R. (eds.), Money, Capital Mobility, and Trade, Essays in Honor of Robert A. Mundell, Cambridge, MA., MITPress. Jones, R.W., Kierzkowski, H. (2005), International fragmentation and the new economic geography, North American Journal of Economics and Finance 16, 1-10. Kohler, W. (2001), A specific-factors view on outsourcing, North American Journal of Economics and Finance 12, 31-53. Kohler, W. (2003), The Distributional Effects of International Fragmentation, German Economic Review 4, 89-120. Kohler, W. (2004a), Aspects of International Fragmentation, Review of International Economics 12, 793-816. Kohler, W. (2004b), International Outsourcing and Factor Prices with Multistage Production, Economic Journal 114, C166C185. Kohler, W. (2007), The Bazaar Effect, Unbundling of Comparative Advantage, and Migration, in: W. Franz, H.J. Ramser, M. Stadler (eds.), Dynamik internationaler Märkte, Tübingen, Mohr-Siebeck, 147-182. Krugman, P. (1995). Growing world trade: Causes and consequences, Brookings Papers on Economic Activity I, 327-377. McLaren, J. (2000), 'Globalisation' and Vertical Structure, American Economic Review 90, 1239-1254. Mann, C.L. (2003), Globalization of IT services and white collar jobs: The next wave of productivity growth, International Eco-
Offshoring: Why Do Stories Differ?
49
nomics Policy Briefs, Institute for International Economics, Washington D.C. Mann, C.L. Kirkegaard, J.F. (2006), Accelerating the Globalization of America: The Role of Information Technology, Washington D.C., Institute for International Economics. Markusen, J.R. (2006), Modeling the Offshoring of White-Collar Services: from comparative advantage to the new theories of trade and FDI, in: B. S. Lael and S. Collins (eds.), Brookings Trade Forum 2005: Offshoring White-Collar Work, Washington, The Brookings Institution, 1-34. Melitz, M. J. (2003), The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity, Econometrica 71 (6), 1695-1725. OECD (2007), Offshoring and Employment: Trends and Impacts, OECD, Paris.
Part 2: The EU and Eastern Europe
Martin Falk and Yvonne Wolfmayr
Home Market Effects of Outward FDI: Evidence Based on Amadeus Firm-Level Data1 Abstract In this paper we investigate substitution possibilities between activities abroad and domestic activities using firm level data on Austrian and German MNEs drawn from the “AMADEUS” database for the period 2000 - 2004. As a measure of domestic activities we use both parent company employment and turnover of parent companies at home. Activities abroad are measured as foreign affiliate employment and foreign affiliate turnover. We also distinguish between manufacturing and non-manufacturing as well as between different host regions (i.e. EU-15 and CEEC). In order to control for the possible impact of outliers we use robust regression techniques. Estimates of the robust regression method and instrumental variable methods indicate that foreign and domestic economic activities and foreign operations are complementary to each other at least for Austrian multinationals. This holds also for foreign operations in CEEC countries. Overall, the resulting effects are very small, however. I. Introduction Since the beginning of the nineties Austria and to a lesser extent Germany have experienced a marked catch–up in cross-border direct investments. For Austria, the stock of outward foreign direct investments relative to GDP has accelerated enormously from only about 3 per cent in 1990 to well over 20 per cent in 2006. The same picture arises from statistics on Austrian inward FDI. However,
1
An earlier version of the paper was presented at the ETSG 2006 conference and at the Int. Economic Workshop "The EU and Emerging Markets: Trade and Investment Relations with Latin America, East Asia and Transition Countries" 2007 at the Vienna University of Economics and Business. Financial support from the Austrian National Bank, Anniversary Fund, Grant No. 11778 is kindly acknowledged.
54
Martin Falk and Yvonne Wolfmayr
during recent years, Austria turned from a net importer into a net exporter of direct investment. Investments in Central and Eastern European countries (CEEC) played a major role in that process. In fact, a significant part of Austrian FDI is nowadays concentrated in the CEEC, accounting for 38 per cent of the total outward investment stocks in 2004 and as much as 72 per cent of the total foreign labour force of Austrian MNEs. This latter figure translates into an employment in Austrian affiliates of 266,000 people, up from about 65,000 in 1994. Within the CEEC region, the largest part of the FDI stock concerns investments in the CEEC-52), which are the countries nearest to the Austrian border. However, since about 2000 there are signals that investor’s interest is shifting to the more distant CEEC, with their share in total foreign affiliate employment increasing from 3 per cent in 1993 to 25 per cent in 2004. In Germany growth of inward and outward FDI has also been very dynamic with affiliate employment of German MNEs roughly doubling over the course of the 1990s. Outward FDI into the CEEC has also played a major role in this process. Overall, however, the CEEC play a lesser role as compared to Austrian FDI. The share of the foreign labour force of German MNEs working in the CEEC amounted to 37 per cent in 2003, roughly half the Austrian share. The growing importance of labour abundant, low-cost locations in Austrian and German outward FDI has spurred increasing public concerns on the possible negative repercussions on the domestic labour market, especially the employment of the low-skilled. The theoretical predictions on the home market effects are far from clear cut and depend on the type and the motive for outward foreign direct investments. While the basic distinction is between vertical (primarily cost-seeking) and horizontal (primarily market-seeking) FDI, with either type, labour market outcomes will further depend on the very specific relationships between the parent company and its foreign affiliates. That is, exactly what parts of the production process are moved across the borders by vertical FDI, (the substitutability or complementarity of offshored production to the activities at home), the overall effect on costs and productivity in the parent company or the scope of the production processes being rep-
2
The group of CEEC-5 includes: Poland, Czech Republic, Slovak Republic, Hungary and Slovenia.
Home Market Effects of Outward FDI
55
licated by the foreign affiliate in the case of horizontal FDI (Molnar et al., 2007). With the large diversity of integration strategies across firms the empirical literature so far also reveals a very heterogeneous picture across countries and industries of the relationship of employment in the parent companies and their foreign affiliates (see Molnar et al., 2007). The diversity of econometric methods applied and the differences in sample coverage make it also hard to discern what exactly drives the differences in the results. In this paper we investigate substitution possibilities between activities abroad and domestic activities using firm level data on Austrian and German MNEs drawn from the “AMADEUS” database for the period 2000 - 2004. Austria’s strong focus on the CEEC as a location of outward FDI makes it a very special case to which we will compare results for Germany. As a measure of domestic activities we use both parent company employment and turnover of parent companies at home. Activities abroad are measured as foreign affiliate employment and foreign affiliate turnover. We also apply instrumental variables estimations to account for the potential endogeneity of the performance of foreign affiliates. In order to control for the possible impact of outliers we use robust regression techniques. II. Related literature The empirical evidence in the related literature yields no conclusive results. For the US, in an early study for the period 1983 - 1992 and a panel of manufacturing multinationals, Brainard and Riker (1997) find a weak, but significant labour substitution effect between parent companies and their affiliates. Their results also show that the substitutability in employment is much stronger among different affiliates of the same MNE that are located in different low cost host countries. Two later papers for the US, both based on BEA’s (Bureau of Economic Analysis) firm level data find a complementary relationship between increases in affiliates’ activities and employment in the parent companies: Hanson et al. (2003) examine MNE data for the two sub-periods 1989 - 1994 and 1994 - 1999 and find a significant positive association between higher sales in foreign affiliates and US parent employment, with stronger results for the period 1989 - 1994. Desai et al. (2005) look at data spanning the period 1982 - 1999 and conclude with a significant positive re-
56
Martin Falk and Yvonne Wolfmayr
lation between employment growth in the foreign affiliates and employment in the parent companies. Both of the latter papers make no distinction between low-wage and high-wage countries. Yet another study for the US (Harrison et al., 2007) using BEA-firm level data, examining simple correlations between US multinational employment at home and abroad for the period 1977 - 1999 finds positive correlations between job growth in foreign affiliates located in high-income countries and employment in the US parent company. The correlation affiliates in low-income countries and the parent company is negative. The latter substitution effects are more pronounced in some of the key sectors such as computers, electronics and transportation. On the basis of survey data on the international expansion of German and Austrian firms, Marin (2004) estimates that since the fall of the iron curtain, German multinationals have shed about 100.000 jobs in Germany in substitution for jobs in the CEEC. For Austria, Marin (2004) calculates that 24.000 jobs were lost due to Austrian FDI in the CEEC in the same period. A recent study by Becker et al. (2005) also derives a substitutive relationship between employment at home and in foreign affiliates in the CEEC using individual data on German and Swedish multinationals and affiliates. According to that study, a drop in wages in foreign affiliates located in the CEEC of 10 per cent - with constant wages in the homeeconomy - leads to a reduction in home-country employment of 0,5 per cent. However, the authors additionally find that the relationship between home and foreign employment more strongly reacts to differences in relative wages within high-wage countries than between Germany and the CEEC. Furthermore, using data for 158 Swiss multinationals Henneberger and Ziegler (1999) find that foreign employment does not affect employment at home. Using a firm-level data for the manufacturing sector for Germany Döhrn (2003) finds that investment abroad has a negative impact on employment at home. Based on firm level data for twelve EU countries for the years 1994 - 1998 extracted from the AMADEUS databank, Cuyvers et al. (2005) report further evidence of employment substitution of outward FDI in the CEEC. Using Amadeus data for Belgium and France, a European Commission (2005) study finds substitution between parent company and affiliate employment in the CEEC, while there is a neutral (Belgium) respectively a complementary (France) employment relationship with affiliates located in EU-15 host countries.
Home Market Effects of Outward FDI
57
These results for EU countries differ from a study by Konings and Murphy (2006) or Murphy (2006), which use firm level data from the AMADEUS databank for European multinationals for the period 1993 - 1998 and find no evidence for substitutive effects between parent employment and the affiliates located in the CEEC. For the manufacturing sector they find a significant negative correlation between parent employment and affiliate employment in the North-EU and EU-15 based foreign affiliates, respectively. These substitution effects are nonexistent for MNEs in the non-manufacturing sector. These results resemble the findings for Austria in Pfaffermayr (2001), based on sector FDI data. For the manufacturing sector and the period 1990 - 1996, in Pfaffermayr (2001), employment in Austrian affiliates located in Western European countries substitutes for employment at home while there is no significant relationship with foreign employment in the CEEC. Using two-way fixed effects panel data methods Falk and Wolfmayr (2007) find a complementary relationship for the services sector, but very weak substitutive effects in the manufacturing sector. The study by Falzoni and Grasseni (2005) for Italian MNEs highlights the importance to account for different firm size. They find a general relocation of jobs in small firms, while for larger Italian firms it is only FDI in Asian countries that exerts a negative impact on home employment. Also for Italy, Barba Navaretti and Castellani (2004) compare the employment performance of purely domestic firms and multinationals (with similar characteristics prior to FDI) and find no significant differences. Molnar et al. (2007) provide an excellent survey on these studies, and highlight this diversity of results which limits the possibility yet to draw firm conclusions on the issue of employment effects of FDI. III.
Preliminary facts on employment in Austrian and German MNEs and their foreign affiliates
The stock of Austrian FDI has accelerated considerably since the early nineties, triggering an enormous catch-up with the scope of internationalisation of production by multinationals in other industrialised countries (see Figure 1). A major driving factor behind this development has been the Austrian involvement in the CEEC. The stock of Austrian FDI in the CEEC has increased tenfold since 1993, exceeding € 18 bn in 2004 or 7 per cent of GDP. In 2005, al-
58
Martin Falk and Yvonne Wolfmayr
most half of the additional €7.5 bn investments were similarly directed into this region. For Germany, there are also strong increases in inward and outward FDI stocks, but the CEEC region is of minor importance accounting for only 3 per cent of GDP (see Figure 2). The importance of the CEEC region as a location for Austrian FDI and the deepness of these cross-border linkages become especially evident in employment figures. In 2004, more than 70 per cent of total affiliate employment was concentrated in the CEEC! This makes Austria a very special case as the respective shares for other European countries are clearly much lower. Even for those countries which are known to have closer ties with the CEECregion. Thus, in Germany, Italy and Finland, the employment shares of FDIs in the CEEC are 37 per cent and 13.5 per cent, respectively (see Figure 3). Mirroring the overall industry pattern of Austrian FDI, the major part of investments in the CEEC is accounted for by the service sector. The sector’s share in total foreign stocks in the CEEC amounted to 80 per cent in 2004. However, the predominance of the services sector shrinks markedly when employment figures are used: The share of the service sector is still high but the share reduces to about 60 per cent of total foreign employment of Austrian MNEs in the CEEC. The services industries also account for the major share in German FDI into the CEEC with about 56 per cent of foreign employees in the German affiliates in the CEEC working in the service sector. Overall employment in Austrian foreign affiliates developed very dynamically over the period 1994 - 2004, increasing by an annual average rate of 13.3 per cent, with affiliate employment in the CEEC growing by an annual average rate of 15.1 per cent. This translates into an overall increase of 264,000 jobs within the period of which 200,000 were created in CEEC affiliates (see Figure 4). Affiliate employment growth was also strong in other locations over the period, slowed down however in the more recent years since 2000, while growth kept strong in the CEEC even in the more recent years. However, there is a clear tendency, especially since 2000, within the CEEC of a shift in FDI and affiliate employment growth away from the traditional markets in the CEEC-5 towards the more distant CEEC. German MNEs employed about 2.6 million workers abroad. Overall, German affiliate employment has been growing at a yearly rate of 3.9 per cent. Mirroring the Austrian data, however, affiliate
Home Market Effects of Outward FDI
59
employment growth has been most dynamic in the CEEC increasing at a yearly rate of 13.4 per cent in the period 1995 to 2003 and creating 600.000 additional jobs in the CEEC (see Figure 5). Figure 1: Austria's inward and outward FDI stocks 22
As percent of GDP
20 18 16 14 12 10
Outward - Total
Inward - Total
8 6 4
Outward - CEEC19
2
Outward - CEE5 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
0
Source: OeNB.
Figure 2: Germany's Inward and outward FDI stocks 35 30
Outward - Total
As percent of GDP
25 20
Inward - Total
15 10 5
Outward - CEEC19
0 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Source: Eurostat, New Cronos.
60
Martin Falk and Yvonne Wolfmayr
Figure 3: Affiliate employment in the CEEC as per cent of total affiliate employment abroad, 2004 (or latest available year) 80
71
70 60 in %
50 37
40 30
17
20
13,5
7
10
1
0 Austria
Germany
Sweden
Italy
United Kingdom
Finland
Source: OeNB, OECD FATS, Eurostat, UNCTAD.
Figure 4: Employment in Austrian foreign affiliates by region 400.000 Rest
350.000 300.000
Other CEEC
36.875
94.341
Persons
250.000 28.651 21.394
200.000 150.000 8.547 100.000
CEEC5
171.837
EU15
67.472
141.012
3.195
50.000
61.890 32.430
57.571
0 1994
Source: OeNB.
2000
2004
Home Market Effects of Outward FDI
61
Figure 5: Employment in German foreign affiliates by region 2.800.000
2.400.000 958.249
960.352
2.000.000
Persons
Rest 1.600.000
834.607 735.333
1.200.000
950.716 CEEC
348.729 800.000
400.000
733.664
836.315
1995
1999
EU15
693.035
0 2003
Source: OECD FATS.
IV.
Database and descriptive statistics
The micro data on Austrian multinationals and its affiliates used in this paper is mainly derived from the “AMADEUS” database of company accounts which is provided by the Bureau Van Dijk. Amadeus covers only European firms and thus limits the information on European affiliates of Austrian and German multinationals. The database has also been used by Cuyvers et al. (2002) and Konings and Murphy (2006) for a panel of European firms to analyse a similar question. From the AMADEUS database we extracted data for all Austrian and German companies holding a minimum share of 10 per cent in a foreign (European) subsidiary. On the basis of information on the parent-affiliate ownership structure all foreign affiliates were identified and linked to the data of the parent company in Austria. For the sample of selected parent and affiliate companies we extracted data on the number of employees, the turnover, the cost of employees, the 4-digit industry NACE-code, and the nationality of the subsidiary for the period 1993 to 2005. Unfortunately we found only limited coverage of the relevant variables for the years 1996 to 1999 and also 2005, so that in the empirical analysis we had to stick to the period 2000 - 2004.
62
Martin Falk and Yvonne Wolfmayr
Table 1: Descriptive statistics
p.a. turnover, median change turnover CEEC, median change turnver non CEEC, median change employment, median change employment CEEC, median change employment non CEEC, median change
Austria Germany parent foreign no. of parent foreign no. of company affiliate obs company affiliates obs 3.9 4.9 404 2.6 3.7 5095 2.5
19.8
99
2.6
14.4
367
4.1
3.3
305
2.6
3.1
4728
0.7
0.6
151
0.0
0.0
3397
0.0
6.6
34
0.0
4.0
240
0.9
0.0
115
0.0
0.0
3156
Source: Own calculations based on AMADEUS.
Table 1 shows summary statistics for the change in parent company's performance and the change in their foreign affiliates. The median employment growth of Austrian parent companies is about 0.7 per cent p. a. compared to 0 per cent for German parent companies. As expected we observe high employment and turnover growth rates in the CEEC subsidiaries. This holds for both Austrian and German multinational enterprises. V. Empirical specifications In the empirical work we relate domestic activities to activities abroad. As measures of domestic activities we use both parent company employment and turnover of parent companies at home and relate these to foreign affiliate employment and affiliate turnover. In specific, we relate the average annual changes between 2000 and 2004 of the respective variables. Expressed in first-differenced logarithmic form we test: 6
' ln PLi
D 0 D1' ln ALi D 2 dat i
¦
4
E j dsec ij
j
D 0 D1' ln AYi D 2dat i
j
Q i (1)
ij
Q i (2)
4
¦ E dsec ¦K dsize j
j
ij
j
6
' ln PYi
¦ K dsize
ij
j
j
where the left-hand side variables PLi and PYi are parent company employment and turnover, respectively. ALi and AYi denote affiliate employment and turnover. dsecij and dsize ij are
Home Market Effects of Outward FDI
63
set of sector and firm size dummy variables referring to the parent company. dat i is the dummy variable for multinationals located in Austria. ' is the difference operator and refers to the average annual change of the variables between 2000 and 2004. Q i is the error term that is assumed mutually independent and normally distributed. In order to test whether the slope parameter differs between Austrian multinationals and German MNEs we introduce an interaction variable that is created by multiplying the dummy variable for Austrian multinationals by the employment change in foreign affiliates. The regression equations can be rewritten as: ' ln PL i
D 0 D 1 ' ln AL i D 2 dat i D 3 dat i ' ln AL i 6
4
j
j
¦ E j dsec ij ¦ K j dsize ij Q i ' ln PLi
D 0 D1' ln ALi D 2 dati D 3dati ' ln AYi 6
4
j
j
¦ E j dsecij ¦K j dsizeij Q i
(3)
(4)
The sign of the coefficient of the interaction term indicates whether the impact of foreign activities is higher (+) or lower (-) for Austrian multinational companies than for German MNEs. With these specifications in long differences, possible fixed effects are eliminated so that we can estimate the models using OLS. However, to control for extreme observations influencing the mean, we apply robust regression techniques. The sign of the coefficients will indicate whether foreign affiliate employment and domestic employment are complements or substitutes. If there is a positive correlation between employment in the parent company and their affiliates in the labour demand equation, then the two factors are complementary to each other. VI.
Results
Table 2 shows the estimation results for the relationship between the turnover of the parent companies and the foreign affiliates. All regressions are performed using the robust regression technique that gives less weight on observations with large residuals. Furthermore, we provide separate estimates by partner region and industry group. We find a significant and positive coefficient on foreign affiliate turnover for the total sample. This means that
64
Martin Falk and Yvonne Wolfmayr
firms that expand their foreign activities simultaneously also expand their domestic operations. Table 2: Robust regression results for the impact of the change of foreign affiliate turnover on parent company turnover Total coef. t 0.005 2.14 0.013 1.26
change in turnover change in turnover *AT dummy variable MNE in Austria (AT) 0.021 firm size employees [75 - 249] (ref cat. <75) 0.001 firm size employees [250 999] 0.009 frim size employees [100010000] -0.001 firm size > 10000 employees -0.016 Other production (ref cat. Investment goods) 0.020 consumer manufacturing -0.039 goods materials manufacturing goods -0.015 wholesale & retail trade -0.019 transport and services -0.014 business services -0.010 constant 0.038 no. of obs 5499
4.44
CEEC coef. t 0.000 -0.03 0.025 1.44 2.76
0.020
3.71
0.17
-0.015 -0.82
0.002
0.42
2.21
-0.030 -1.89
0.012
2.89
-0.21 -4.59
0.011 0.76 -0.031 -2.49
4.22
-0.015 -0.84
-5.63 -3.70 -4.14 -3.05 -2.80 9.43
0.033
Non-CEEC coef. t 0.006 2.19 -0.006 -0.32
-0.045 -0.010 -0.033 -0.029 -0.022 0.054 466
-2.18 -0.69 -2.00 -1.44 -1.70 3.62
-0.002 -0.43 -0.014 -3.82 0.022
4.49
-0.037 -0.016 -0.018 -0.012 -0.010 0.037 5033
-5.05 -3.69 -3.66 -2.67 -2.57 8.63
The coefficient of the interaction term between Austrian multinationals and foreign affiliate turnover growth is not significantly different from zero, indicating that the effect is not different for Austrian multinationals. Turning to the subsample of multinationals and their affiliates in CEEC we find that the expansion abroad does not affect the growth of turnover of the parent company. The coefficient of the interaction term shows that turnover growth of Austrian affiliates located in the CEEC is also complementary to changes in turnover of the parent company, but that the coefficient is not significant at the 10 per cent level.
Home Market Effects of Outward FDI
65
Table 3: Robust regression results for the impact of the change of foreign affiliate employment on parent company employment Total CEEC coef. t coef. t -0.001 -0.50 0.015 0.80 0.048 1.77 0.036 1.98
change in employment change in employment*AT dummy variable MNE in Austria (AT) 0.005 firm size employees [75 - 249] (ref cat. <75) 0.002 firm size employees [250 999] -0.013 frim size employees [100010000] -0.021 firm size > 10000 employees -0.031 Other production (ref cat. Investment goods) -0.015 consumer manufacturing goods 0.001 materials manufacturing goods -0.013 wholesale & retail trade -0.003 transport and services 0.004 business services 0.001 constant 0.014 no. of obs 3547
Non-CEEC coef. t -0.001 -0.44 0.015 0.42
1.02 0.009
-0.87
0.009 1.60
0.68 0.008
1.98
0.001 0.21
-4.84 0.008
0.97
-0.016 -5.36
-7.31 0.007 -8.19 0.010
1.34 -1.46
-0.025 -7.99 -0.034 -8.28
-3.86 0.18 -3.91 -0.99 1.01 0.35 4.50
-1.74 1.78 -1.73 1.12 1.16 -1.50 -1.26
-0.016 -0.002 -0.013 -0.004 0.004 0.001 0.017 3271
0.010 0.012 0.008 0.009 0.010 0.008 0.008 276
-3.84 -0.35 -3.65 -1.04 0.99 0.33 5.07
Turning to results for employment changes we find that employment changes in Austrian affiliates located in CEEC are significantly positively related to the employment changes in the parent companies (see Table 3). For German multinationals we find that foreign employment does not have any impact on parent company employment. Table 4 shows the estimation results of the impact of foreign company employment on parent company employment using instrumental variable regressions. The first stage regression includes the initial log of foreign affiliate employment multiplied by the dummy variable for Austria, as well as sector and size dummy variables. The estimated effect for foreign affiliate employment is somewhat larger than that of the OLS regressions presented in Table 3.
66
Martin Falk and Yvonne Wolfmayr
Table 4: IV regression results for the impact of the change of foreign affiliate employment on parent company employment Total coef. t change in employment change in employment*AT dummy variable MNE in Austria (AT) firm size employees [75 - 249] (ref cat. <75) firm size employees [250 999] frim size employees [100010000] firm size > 10000 employees Other production (ref cat. Investment goods) consumer manufacturing goods materials manufacturing goods wholesale & retail trade transport and services business services constant no. of obs
CEEC coef. t
Non CEEC coef. t
0.127 0.055
0.16 -0.262 -0.24 0.173 2.87 0.248 2.18 0.047
0.15 2.42
-0.093
-1.73 -0.167 -1.16 -0.075
-1.30
-0.060
-2.65 0.039
-2.80
0.43 -0.065
-0.243 -12.06 -0.259 -2.91 -0.241 -11.56 -0.249 -11.76 -0.278 -3.52 -0.247 -11.21 -0.304 -10.89 -0.323 -3.16 -0.301 -10.39 0.129
4.59 0.241
-0.065 -1.68 -0.059 -2.46 -0.051 -2.01 -0.063 -2.49 -0.039 -1.83 0.256 11.44 3547
-0.100 -0.035 -0.060 -0.094 0.055 0.250 276
2.28 0.119 -0.77 -0.39 -0.65 -0.84 0.65 3.11
-0.064 -0.063 -0.049 -0.060 -0.045 0.255 3271
4.07 -1.57 -2.50 -1.87 -2.32 -2.02 10.96
VII. Conclusions and future work
The paper gives new evidence for the home country effects of foreign activities. Based on the AMADEAUS firm-level database for Austria we investigate the substitution possibilities between foreign affiliate activities and domestic operations measured as the change in turnover and employment between the period 2000 and 2004. In particular we distinguish between manufacturing and nonmanufacturing and between different locations. We find that foreign activities do not have a negative impact on employment and turnover of the parent company. If anything, the analysis for Austrian multinational companies reveals a small positive relation between employment change of the parent company and employment change of the foreign affiliates. These results also hold for the impact of parent company employment in
Home Market Effects of Outward FDI
67
the CEEC region. For German multinational enterprises we do not find any significant effects. Overall the results are consistent with Desai et al. (2007) who find firms that expand their foreign activities simultaneously also expand their domestic operations, such as turnover or employment. With regard to future research, two directions appear to be promising. First, research can be extended by applying a similar model to other home countries. In particular, one can extend the analysis by including other countries e.g. Italy, Germany and Sweden, countries with a high level of outward FDI activities in the CEEC. Second, since the explanatory power of the regression equations are quite low it will be necessary to include additional variables into the equations. For instance, wages and output of the parent company should enter the labour demand equations. However, a more detailed analysis is limited due to the large number of missing values in the AMADEUS database.
References
Barba Navaretti, G., Castellani, D. (2004), Investment Abroad and Performance at Home: Evidence from Italian Multinationals, CEPR Discussion Paper no. 4284, CEPR, London. Becker, S., Ekholm, K., Jäckle, R., Mündler, M. (2005), Location Choice and Employment Decisions: A Comparison of German and Swedish Multinationals, Review of World Economics (Weltwirtschaftliches Archiv) 127(4), 693-731. Brainard, L. S., Riker, D. A. (1997), Are U.S. multinationals exporting US Jobs?, NBER Working Paper no. 5958. Cuyvers, L., Dumont, M., Rayp, G., Stevens, K. (2005), Home Employment Effects of EU Firms Activities in Central and Eastern European countries, Open Economies Review 16, 153-174. Desai, M. A., Foley, C. F., Hines, J. R. (2005), Foreign Direct Investment and Domestic Economic Activity, NBER Working Paper no. 11717. Döhrn, R. (2003), Investment Abroad and Home Employment in the German Manufacturing Sector, in: Frenkel, M., Stadtmann,
68
Martin Falk and Yvonne Wolfmayr
G. (eds.), Foreign Direct Investment, Theory, Empirical Evidence and Policy Implications, INFER Studies 9, VWF, Berlin, 1-16. European Commission (2005), The Adjustment Challenge in the Labour Market, The EU Economy 2005 Review, Commission of the European Communities, Brussels. Falk, M., Wolfmayr, Y. (2007), Austrian FDI in Central-Eastern Europe and Employment in the Home Market, Austrian Institute for Economic Research mimeo. Falzoni, A. M., Grasseni, M. (2005), Home Country Effects of Investing Abroad: Evidence From Quantile Regressions, CESPRI Working Paper no. 170, University of Bocconi, Milano. Hanson, G. H., Mataloni M. J., Slaughter, M. J. (2003), Expansion Abroad and the Domestic Operations of U.S. Multinational Firms, Tuck Business School, Dartmouth Working Paper. Harrison, A. E., McMillan, M. S., Null, C. (2007), US Multinational Activity Abroad and US Jobs: Substitutes or Complements?, Industrial Relations: A Journal of Economy and Society 46(2), 347-365. Henneberger, F., Ziegler, A. (1999), Arbeitsplatz-Export durch Direktinvestitionen? Empirische Evidenz für die Schweiz 19901996, Jahrbuch für Wirtschaftswissenschaften 50, 136-155. Konings, J., Murphy, A. P. (2006), Do Multinational Enterprises Relocate Employment to Low Wage Regions? Evidence from European Multinationals, Review of World Economics 142 (2), 267-286. Marin, D. (2004), A Nation of Poets and Thinkers – Less So With Eastern Enlargement? Austria and Germany, CEPR Discussion Papers no. 4358, CEPR, London. Molnar, M., Pain, N., Taglioni, D. (2007), The Internationalisation of Production, International Outsourcing and Employment in the OECD, Economic Department Working paper no. 561, OECD, Paris. Murphy, A. (2006), An Assessment of the Implications of EU Enlargement for Foreign Direct Investment and Jobs in Ireland,
Home Market Effects of Outward FDI
69
The Central Bank & Financial Services Authority of Ireland, Quarterly Bulletin 1/2006, 105-118. Pfaffermayr, M. (2001), Employment in domestic plants and foreign affiliates: a note on the elasticity of substitution Weltwirtschaftliches Archiv 137 (2), 347-364.
Özlem Onaran1
The Effect of Trade and FDI on Employment in Central and Eastern European Countries: A Country-Specific Panel Data Analysis for the Manufacturing Industry Abstract This paper analyzes the labor demand based on panel data of manufacturing industry from Central and Eastern European Countries and discusses the effect of domestic factors (wages and output) and international factors (trade and FDI) on employment in the post-transition period. The findings indicate that employment only responds to wages in 50 per cent of the cases. The output elasticity of labor demand is mostly positive, but low, with a number of cases where employment is completely de-linked from output. An impressive speed of integration to the European economic sphere through FDI and international trade has not prevented job losses in the manufacturing industry. While there are very few cases of positive effects, insignificant effects of trade and FDI dominate, some evidence of negative effects appears as well. I. Introduction This paper aims at exploring the development of employment in Central and Eastern European Countries (CEECs) in the post-transition period and the employment effects of integration into the world economy on employment, in particular with the European 1
This is a shortened version of Onaran Ö., "Jobless growth in Central and Eastern European Countries: A country-sepcific panel data analysis for the manufacturing industry," EASTERN EUROPEAN ECONOMICS 46 (4), July-August 2008. The author is grateful to Paul Ramskogler for excellent research assistance, to an anonymous referee, Engelbert Stockhammer, and Markus Leibrecht for fruitful discussions, and to the Vienna Institute of International Studies, and in particular to Monika Schwarzhappel, and Renate Prasch for their valuable support with data. Needless to say, all remaining errors are mine.
72
Özlem Onaran
economic sphere through trade openness and FDI. In the early 1990s, during the initial phase of transition, CEECs faced a severe recession due to both supply and demand shocks as well as major institutional changes. The adjustment in labor markets to these shocks took place under the starting conditions of high levels of disguised unemployment along with high labor force participation rates in the centrally planned economies due to the target of full employment (Brada, 1989; Kornai, 1995; Blanchard, 1998). Nevertheless, in the early phase of transition the reduction in output was much more pronounced than the increase in unemployment. Izyumov and Vahaly (2002) show that the linkages between unemployment change and output were highly unstable during the early transition period. The political concerns about unemployment, the preservation of soft budget constraints in many state owned firms, low labor mobility particularly due to firm-specific non-wage benefits or infra-structure problems are cited as some reasons explaining this inertia. The transition crisis was replaced by a recovery in output starting in the second half of the 1990s in the Visegard Countries and Slovenia and in the late 1990s in the Baltic States and Bulgaria and Romania, but as market transition matured dramatic changes in the sectoral employment structure and wages emerged in the CEECs (Havlik and Landesmann, 2005; Boeri and Terrell, 2002). In general, compared to the pre-transition era there has been a sharp contraction in employment, an increase in open unemployment, a massive exit from the labor market, and only moderate job creation. One question is how much of that negative development in employment in the post-transition era can be related to the previous labor hoarding. While earlier research on “idle employment” in the CEECs indicates a continuation of the problem (eg. Kajzer, 1995; Jackman, 1994) or even an increase in “overemployment” or “labor hoarding” (e.g. Gora, 1995) during the early transition era, later studies find out that employment became much more responsive to recessions after the mid 1990s, indicating that the labor hoarding problem of the previous phase had already started to be reversed (Boeri and Garibaldi, 2006; Basu et al., 2005); also firms in the transition economies started to impose hard budget constraints on each other later in the transition phase (Schaffer, 1998). Basu et al. (2005) estimate employment elasticities with respect to sales based on firm level data and find that Hungary had already high elasticities at the beginning of the transition phase; Poland went into the
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
73
transition less reformed but achieved high elasticities in the transition years; and the Czech Republic and Slovakia also rapidly reached high elasticities, although they started from employment regimes, which were rather unresponsive to sales. Furthermore, the evidence presented by Basu et al. (2005) does not support the hypothesis that State Owned Enterprises responded less flexibly to changes in sales. Based on a panel data analysis for the aggregate economy, Boeri and Garibaldi (2006) show that in the aftermath of 1996, recession periods led to significant job destruction, whereas expansions in GDP did not lead to statistically significant job creation in the CEE-10. Indeed high rates of output growth in the CEECs in the post-recession era generated fewer jobs than stagnation in the other countries of the EU (Boeri and Garibaldi, 2006). Izyumov and Vahaly (2002) find a lower Okun’s coefficient of 0.526 (effect of GDP growth on the change in unemployment) in the 10 CEECs in the post-recession era of 1995-2000 compared to the coefficient for EU15 (-0.799). Based on the empirical evidence in Basu et al. (2005) and Boeri and Garibaldi (2006) that labor hoarding had already started to be reversed during the transition era, the continuation of the problem of slow employment growth more than a decade after the starting of a major processes of privatization and structural change is worth further analysis going beyond the old over employment problems of the planned economies. Lehmann (1995) argues that severe and persistent shortages in capital and managerial ability may result in keeping labor demand weak in the medium term. OECD (2005) points out that the large negative structural shocks in the CEECs, such as those associated with opening economies to trading at world prices have resulted in a substantial increase in unemployment that persists for a considerable period of time. It is one of the concerns of this study to explore the link between employment and output particularly for the period of the 2000s – a time when the countries had achieved a long way in terms of integration into the world economy through market mechanisms. The situation in manufacturing employment is even more dramatic. It decreased in all countries not only in the first period of transition recession, but also in the post-recession period. In general, the jobs created in services have off-set the negative effects of the major downsizing in the manufacturing industry, but even during the uninterrupted growth years of 2000s new service jobs have just sufficed to generate stagnation in total employment
74
Özlem Onaran
(Hungary, Czech Republic, Slovakia, Lithuania, Bulgaria), or in some cases could not even compensate for the job losses in manufacturing (Poland, Romania); only in Slovenia, Estonia, and Latvia showed a modest increase in employment. Table 1 shows the annual average growth rates of GDP and total employment in the period of 2000-2005.2 Table 1: Growth in GDP and employment in CEECs (2000-2005 period average, in per cent) GDP Employment Czech R. 3.6 0.1 Hungary 4.1 0.4 Poland 3.1 -0.4 Slovakia 4.3 0.6 Slovenia 3.4 1.2 Estonia 7.4 0.8 Latvia 7.9 1.1 Lithuania 6.9 0.2 5.0 0.6 Bulgaria Romania 5.0 -0.3 Source: Vienna Institute for International Economic Studies.
Another important concern is about the quality of the jobs created in the service sector. Although the shift in employment from industry towards services is a pattern, which goes along with improvements in productivity, and can be observed in developed countries as well, Reinert and Kattel (2004) point out that the type of deindustrialization in the CEECs is qualitatively very different from the slow ‘de-industrialization’ of high-income countries, which upgrade into a knowledge-intensive service sector; in contrast the service jobs created in the CEECs are mostly low – skilled and low-paid jobs. A further controversial fact is that rapid improvements in exports and foreign direct investment have so far not been able to reverse the stagnation in aggregate employment or the decline in manufacturing employment in the CEECs. Hunya and Geishecker (2005) provide evidence that the nature of FDI flows can account 2
The figures for 2006, which became available after writing this report indicate stronger employment growth in total employment in 2006 compared to 2005 particularly in Poland, Slovakia, and Bulgaria.
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
75
for this development to some extent. About half of the FDI in the New Member States between 1990 and 1998 was in the form of privatization-related acquisitions, and the restructuring of the former state-owned enterprises led to massive labor shedding. In later years, especially in manufacturing, most of the new FDI has been investment in new assets; however even then although new capacities usually increased employment, technological progress also led to lay-offs simultaneously. Moreover, most of the greenfield jobs have been created in the service sector such as banking, retail and real estate. Irrespective of the initial method of entry, FDI is now increasingly taking the form of reinvestment of profits, the results of which are yet to be seen. Apart from the direct effects, indirect negative effects of FDI are also observed (Hunya and Geishecker, 2005): jobs were destructed through negative spillovers as foreign investors replaced traditional domestic suppliers by imports or domestic firms disappeared or downsized due to intensified competition of larger and technologically more advanced subsidiaries of multi-national enterprises. Overall Hunya and Geishecker (2005) find that domestically-owned manufacturing companies reduced the number of employed while foreign-owned enterprises expanded that number. In this study, we shall estimate the effects of FDI and foreign trade on sectoral employment in order to shed light on the aggregate direct and indirect job creation and destruction effects. Regarding the role of labor market institutions in determining labor demand, many indicators show that the newly formed labor markets in the CEECs are rather flexible. Thus wage or employment rigidity does not seem to be the reason behind the disappointing employment performance in the 2000s. Based on panel data estimation of wage bargaining equations for the sub-sectors of manufacturing in the CEECs, Onaran and Stockhammer (2008) find that wages are highly flexible with respect to unemployment. Boeri and Garibaldi (2006) report that wage floors in the New Member States (NMS) are often not binding, and are rarely enforced in the private sector; the ratio of minimum wage to the average wage is around 30 per cent compared to a ratio of 50 per cent on average in EU15. Also collective bargaining coverage rates are very low compared to EU-15, although union density rates are more comparable (Boeri and Garibaldi, 2006). Regarding employment flexibility, Hungary, the Czech Republic, and Slovakia are ranked in the more flexible half of the OECD countries according to the Index of Rigidity of
76
Özlem Onaran
Employment Protection Legislation of OECD (2004). The Employment Rigidity Index in World Bank’s Doing Business Report (2006) ranks the four OECD members in CEE (Czech Republic, Poland, Hungary, and Slovakia, the first being the most flexible) at a level between fifth to ninth among 20 countries, where Ireland is ranked the sixth. The analysis in this paper covers only the manufacturing industry due to data limitations; nevertheless manufacturing industry is rather significant for the economy, accounting for an important part of total GDP and employment, almost half of the FDI, and almost all of the merchandise exports and imports. Also manufacturing employment has been hit hardest throughout the process of transition, and it has not recovered since then. The paper first estimates a labor demand equation for the sub-sectors of manufacturing industry in order to test the effect of output and wages on employment. Second, after controlling for wage and output, the paper tests whether integration to the world economy via foreign trade and FDI improved the employment creation capacity of the manufacturing industry, or quite on the contrary, whether intense competitive pressures have led to a downward shift of the labor demand curve. The labor demand equation is estimated for each country separately based on panel data of manufacturing industry, supplied by the Vienna Institute of International Studies. Due to data availability for trade and FDI at a sectoral level the period of analysis is the later stages of post-transition recovery. Thus, the severe contractions in economic activity in the early and mid 1990s, as well as the possible continuity of the post-recession adjustment process in terms of reducing over employment do not distort the estimations. The countries explored are the eight CEECs, for which data is available - the Czech Republic, Hungary, Poland, Slovakia, Slovenia, Lithuania, Bulgaria, and Romania. The other two Baltic countries, Estonia and Latvia, are left out of analysis due to data problems regarding changing sectoral classification through time. The paper consists of five sections, including this introductory one. Section II presents the labor demand model. Section III discusses the descriptive statistics of the working sample. Section IV summarizes the estimation results, and Section V concludes. II. Labor demand model The analysis in the paper is based on a labor demand equation, derived from a production function following the methodology used
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
77
in Greenaway et al. (1999) and Hine and Wright (1997) for the UK, and Milner and Wright (1998) for Mauritius. This methodology was also used by Stehrer (2004) for analysing employment effects of trade in OECD countries. The analysis is based on a fairly simple model of a profit maximising firm with a Cobb-Douglas production function. In this model, for the sake of simplicity, the cost of capital is supposed to vary only over time, assuming perfect capital markets; thus its variation is captured by time dummies at the stage of estimation (Milner and Wright, 1998). Although this is a strong assumption, it is convenient in our case where there is no reliable data about the costs of capital, and interest rates are not a good proxy. Also it is not convenient to assume a priori that capital is a substitute for labour, rather it can be a complementary input, or it is also possible that the cost of capital can have no direct impact on the demand for labour, since technology is fixed in the short-run. This would also be the case if the firm has excess capacity. Additionally the technical efficiency of production, A, which is part of the constant term, c0, is expected to increase over time, and this effect will be captured by time dummies. At this stage in order to test the effects of openness on labor demand, technical efficiency is modelled as a function of international trade, hypothesizing trade induced technological change (Greenaway et al., 1999). In Milner and Wright (1998) and Hine and Wright (1997) also discuss expected effects of trade liberalization due to the changes in the labor intensity of production as a response to the comparative advantage of the country. The effects of trade through these different channels are discussed in more detail below. As an extension to these models we also integrate the effect of FDI on the technical efficiency and labor intensity of production. Thus the following function forms the basis of the estimations:
ln Li
f Ei ,Dt , lnQi , lnWRi , FDIQi , XQi , MQi
(1)
where lnLi, lnQi, lnWRi, FDIQi, XQi and MQi are the employment (in logs), real output (in logs), real wage (in logs), inward FDI stock/output, exports/output and imports/output in sector i respectively. E i is a sector specific coefficient. t is the time dummy, capturing the changes in capital costs and other time specific shocks as well as technical change not captured by international trade and FDI. Other factors such as employment taxes or institutional factors
78
Özlem Onaran
like employment legislation that may affect labor demand are captured also by the time dummy, since the analysis of these effects, albeit interesting, are outside the scope of this paper. The effects of international trade and capital flows on employment vary among economic theories. Based on the Heckscher-Ohlin theorem, traditional trade theory indicates that in a labor and particularly unskilled labor abundant developing country, after trade liberalization the employment of unskilled labor, and in general employment in export sectors increases due to the comparative advantage of the economy in more labor intensive sectors, whereas the employment of skilled labor or certain groups of labor specialized in import-competing industries may fall in spite of aggregate welfare gains. These effects are generated by changes in relative prices of exportables and importables after trade liberalization. Therefore, in order to test the trade effects based on trade theory, one has to test the effect of trade on relative prices of exportables vs. importables; and then through the price channel, the effects on the demand for factors used more intensively in the exporting sectors (in the case of developing countries labor) vs. the import competing sectors can be estimated. Indeed the methodological debates to test the effect of trade on employment has a long history by now, which also has its roots in differences in theoretical approaches. Apart from trade theory, labor economics approaches based on factor content analysis evaluate the effects of trade with regard to shifting labor demand in response to exports, which is a source of demand, and to imports, which is a reduction in demand (e.g. Wood, 1994). Thus exports increase employment, whereas imports decrease it.3 However this methodology is criticized by trade theoreticians, since it takes the changes in trade volumes and not relative prices as the starting point. Finally different from the trade theory or labor economics approaches, based on a microeconomic perspective, trade not only shifts the demand schedules, but may also bring international competitive pressure, which may lead to trade induced technological change or efficiency gains (Greenaway et al., 1999; Rodrik, 1997). Particu3
However, if imports are no substitutes of domestically produced goods, but mostly complementary input goods, that are not being produced domestically, this negative effect will not be observed, or even a positive effect is possible (Onaran, 2001).
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
79
larly in the empirical literature for the advanced countries, there is an increasing consensus that the magnitude of trade flows is far too low to account for changes in labor market outcomes, but defensive innovation stimulated by international competition may have an indirect negative effect on employment (e.g. Stehrer, 2004; Greenaway et al., 1999). Similarly, stylized facts from many labor abundant developing countries indicate that an impressive performance in terms of increasing exports has not brought with it parallel increases in employment contrary to the expectations (Horton, et al., 1994; Amsden and Hoeven, 1996; Onaran, 2001; Pollin et al., 2004; Ghosh, 2005). In this paper, following Hine and Wright (1997) and Greenaway et al. (1999) regression analysis is used to test the labor demand effects of trade due to both changing labor intensity of production for a given level of output and trade induced labor saving effects. In that sense, the estimation methodology used here is not a direct test of the Hecksher-Ohlin framework. However, since mainstream policy is based on a narrower reading of the Hecksher-Ohlin framework in formulating an optimistic expectation in terms of job creation due to increased trade liberalization and volume, this can be regarded as a test of these policy expectations as well. In the case of foreign direct investment, opposing effects are at work once again. If multinational enterprises are bringing a labor saving, more capital intensive technology, it would lead to a negative effect on employment even in export-oriented sectors. Moreover, the nature of FDI, whether it is in the form of greenfield FDI, which creates new production capacity, or brownfield FDI, which involves mergers and acquisitions, matters. If FDI is mostly through mergers and acquisitions, rather than a genuine long term investment, then labor shedding and downsizing may dominate positive employment creation capacity. Second, even when the positive firm level effects of FDI are realized, the spill-over effects can be quite limited, which may lead to a dual economy, without any linkages to the domestic economy, and may even create negative effects on sector wide employment through competitive pressures on domestic firms. Mencinger (2003) reports that multinational enterprises contributed more to imports than exports since they use their international suppliers rather than the domestic firms for intermediate inputs; and the spillovers from single firms to the sector do not seem to be sufficiently strong to increase growth. FDI could also force small local competitors out of business. In this case
80
Özlem Onaran
FDI can favor a particular group of skilled labor, while the unskilled labor’s employment may decline. Thus it is an empirical issue whether the positive or negative effects dominate at the sectoral level. In the empirical analysis the effects on different sectors are also distinguished by grouping sectors as high skill and medium-low skill. In traditional trade theory the expected positive export effect is mostly limited to medium-low skill sectors in countries where high skilled labor is scarce. According to the trade induced efficiency theories, the competitive pressures can be higher in highskill industries in the CEECs, leading to a higher negative effect of trade. Differently, there are studies that emphasize the skill-bias in international trade and capital flows, which expect a positive effect of openness on employment in the high-skill sectors, and a negative effect on medium-low skill sectors even in countries, where skilled labor is relatively scarce. Feenstra and Hanson (1996) illustrate this result based on the case of US and Mexico trade, where skill intensity of production is increasing in both the skilled labor abundant North and the skilled labor scarce South. III.
Descriptive statistics
This section presents the stylized facts of our working sample for each country for the total manufacturing industry as well as high, medium, and low-skill sector groups. Some notes about the data should be mentioned here. FDI data is only available at the one-digit level, i.e. corresponding to 14 sectors, and covers only 1997/1999-2004 for most countries. Furthermore, for Romania and Bulgaria there is no sector specific FDI time series data. The rest of the data is at the two-digit sectoral classification for Hungary, Poland, Slovenia, and Lithuania, and at the one-digit level for the other countries.4 Regarding trade, the focus is on trade of the CEECs with EU 15, since that reflects a pattern of international division of labor and specialization in trade between the center and the periphery. Trade with EU15 reflects roughly 50-75 per cent of the foreign trade volume of the CEECs. Trade data based on the records of the member states themselves is available only from 1999 onwards. 4
The skill classification is based on Landesmann et al., 2004. See Onaran (2008) for details.
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
81
Table 2 shows for each country the average annual per cent change (compound average) in employment, real output, real wage, productivity, and the averages of export/output, import/output, and FDI (inward stock)/output ratios for the pools of high, medium, and low-skill industries, and aggregate manufacturing industry during the period 1999-2004. Although the data suggests mixed performance across countries as well as industries within the same country, it seems to be clear: There is a relatively strong opening up of the economies, improved exports in manufacturing and strong FDI inflows do not go along with a parallel strong improvement in employment creation capacity in manufacturing industry, and productivity increases seem to be based on downsizing and labor saving to a significant extent. The employment trends the transition period determined by rationalization and labor saving seem to be continuing. During 1999-2004 employment has decreased in most of the sectors, with the exception of the medium-skill sectors in Slovenia, Slovakia, the Czech Republic, and Lithuania, and the low-skill sectors in Bulgaria. At a more detailed sectoral level (not reported in Table 2), one observes that employment has increased only in a minority of the industries, which in the case of Hungary correspond to the strong exporting and growing sectors. In Bulgaria and Romania employment decreased in almost all sectors. In Slovenia output has also decreased in half or more than half of the sectors, leading to an expectedly poor employment performance as well. In Hungary, the Czech Republic, and Slovakia the sectors, where employment has increased are mostly the skilled sectors, whereas in Poland, and Lithuania they were the unskilled sectors, and in Slovenia it is more mixed. These developments contrast sharply with the increases in output in almost all sectors in the same period, with the exception of high and low-skill sectors in Slovenia. Output has been increasing with quite high rates in some of the leading exporter industries like electronics or automotive industry. In the meantime real wages significantly lag behind productivity in aggregate manufacturing industry as well as in almost all sector groups in all countries during the post-recession era, with the single exception of low-skill industries in Slovenia, In Poland, Hungary, and Slovakia the difference is most pronounced. However,5 this 5
However, it should be noted that unit labor costs in Euros have been increasing in most cases during this period in spite of the positive
82
Özlem Onaran
decline in unit labor costs due to high productivity and low wage increases has not sufficed to stimulate employment. Table 2: Developments in manufacturing industry (to be cont.) Average annual change 1999-2004 in per cent Employment Real output Real wage Productivity HUN total manufact. high-skill medium-skill low-skill POL total manufact. high-skill medium-skill low-skill SLO total manufact. high-skill medium-skill low-skill LIT*** total manufact. high-skill medium-skill low-skill SLK total manufact. high-skill medium-skill low-skill CZ total manufact. high-skill medium-skill low-skill BU total manufact. high-skill medium-skill low-skill RO total manufact. high-skill medium-skill low-skill
-1.04 -0.31 -0.28 -4.25
9.56 8.22 6.15 1.49
5.03 5.39 5.06 4.15
10.71 8.66 6.41 5.93
-3.30 -3.64 -1.22 -4.21
6.58 6.68 7.79 3.64
1.00 1.78 0.20 0.31
10.21 10.73 9.12 8.18
0.01 -7.43 1.24 -1.88
3.82 -2.04 3.02 -3.18
2.46 2.75 1.45 1.90
3.81 6.17 1.76 -1.34
-1.74 -4.70 1.20 -1.13
10.28 9.57 16.60 12.50
0.84 2.51 1.37 1.69
12.23 14.92 15.21 14.12
-0.81 -2.17 1.18 -1.53
8.94 9.81 7.22 8.11
1.70 2.94 1.17 1.00
9.83 12.37 5.92 9.74
-1.08 -0.13 0.87 -6.32
6.64 8.11 6.55 0.24
3.97 3.80 3.26 3.60
7.80 8.29 5.66 6.80
-0.81 -6.12 -1.68 1.78
10.16 8.03 12.10 14.20
-0.51 1.97 -0.77 1.15
11.06 15.20 14.30 12.22
-2.11 -3.75 -2.35 -0.58
7.07 6.45 9.51 6.15
5.74 7.15 6.08 4.69
9.38 10.81 12.17 6.91
difference between productivity and wage increases because of currency appreciation.
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
83
Table 2 (continued): Developments in manufacturing industry Annual Average (1999-2004*) Export/output Import/output (to EU15) (from EU15) FDI stock/output** HUN total manufact. high-skill medium-skill low-skill POL total manufact. high-skill medium-skill low-skill SLO total manufact. high-skill medium-skill low-skill LIT*** total manufact. high-skill medium-skill low-skill SLK total manufact. high-skill medium-skill low-skill CZ total manufact. high-skill medium-skill low-skill BU total manufact. high-skill medium-skill low-skill RO total manufact. high-skill medium-skill low-skill
0.539 0.495 0.279 0.939
0.482 0.601 0.437 0.752
0.260 0.267 0.229 0.255
0.248 0.279 0.175 0.360
0.319 0.497 0.245 0.290
0.142 0.124 0.126 0.238
0.481 0.507 0.341 0.595
0.588 0.655 0.449 0.533
0.112 0.161 0.133 0.082
0.327 0.376 0.246 0.453
0.438 1.293 0.644 0.268
0.140 0.161 0.153 0.106
0.413 0.472 0.225 0.576
0.384 0.544 0.298 0.335
0.134 0.103 0.183 0.099
0.492 0.530 0.321 0.630
0.479 0.635 0.385 0.492
0.221 0.225 0.193 0.215
0.354 0.233 0.300 0.944
0.103 0.875 0.657 0.470
0.312 0.254 0.128 1.051
0.103 0.875 0.657 0.470
Notes: *For Bulgaria and Romania 1999-2003; **FDI data for Czech Rep., Poland, Slovenia covers 1999-2003; ***For Lithuania 1999-2001.
84
Özlem Onaran
In the same period, the trade and FDI record of all countries is impressive. Regarding the pattern of trade, in all countries the highest export/output ratio (destinated to EU markets) is observed in low-skill sectors, and the highest import/output (from EU) ratio in high-skill sectors (except Hungary); however usually import ratios in other sectors are also quite high. These stylized facts indicate the existence of intra-industry trade in all sectors along with some comparative advantage in low-skill sectors. Initially, the CEECs started with a profile typical of less developed economies in terms of the concentration of their exports to EU 15 in labor intensive sectors. But over time trade positions in capital, R&D and skill intensive industries improved in the more advanced CEECs, with Hungary experiencing the most remarkable change, while Romania and Bulgaria stayed remarkably different from the CEE-5 (Havlik et al. 2001). Nevertheless, since then there is a trade deficit with the EU in high and medium-skill sectors in all countries and a trade surplus in low-skill industries. In terms of the stock of FDI as a ratio to output, in the Czech Republic and Hungary, all sectors, including some high-skill sectors like electronics, are important recipients. In Slovenia and Lithuania high-skill sectors followed by the medium-skill have the highest FDI intensities. In Slovakia medium-skill industries are the leading FDI recipients. Poland is the only country, which has the highest FDI stock/output ratio in low-skill industries. Regarding the relative importance of firms with 10 per cent or more foreign shareholders, foreign penetration of the domestic economy is highest in manufacturing in Hungary, with 45 per cent of the workforce employed in foreign subsidiaries as of 2001; this is followed by the Czech Republic, Slovakia and Poland, with a penetration rate of around 35 per cent (Hunya and Geishecker 2005). Romania shows a lower foreign penetration (close to 30.7 per cent). Slovenia has the lowest foreign penetration in manufacturing due to the domestic economic policy which has not encouraged capital inflows since Slovenian companies had been already at the outset of transformation integrated internationally and were largely competitive which was also supported by a stable real exchange rate (Hunya and Geishecker, 2005). In general, foreign affiliates are characterized by higher labor productivity due to more up-to-date technology than domestic companies as well as narrower specialization on assembly and component production, i.e. benefiting from economies of scale (Hunya and Geishecker, 2005).
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
85
IV. Estimation results
In order to analyze the determinants of employment, country specific panel data estimations are used. The panel data technique addresses the research questions based on variations both over time and across sectors within each country, while allowing for cross country heterogeneity. The other advantage of panel data is the technical capacity, which makes empirical tests possible with a database of relatively short time dimension. Both static and dynamic versions are estimated, and depending on the significance of the lagged dependent variable we decided between different specifications. Additionally, the real wage is treated as an endogenous variable. The estimations are made first for the total pool of all the sectors, and then for separate pools of high-skill (HS) and medium low-skill (MLS) sectors. In the following we summarize the results. The tables and further details are in Onaran (2008). According to the results of the aggregate pool, in six countries output has a significant positive effect on employment. Only in Lithuania and Romania the effect is insignificant. These results differ from the insignificant elasticities found in Boeri and Garibaldi (2006) for the positive growth years during the period of 1996-2002 based on panel data estimations for the aggregate economy. Thus firms were not completely reluctant to hire during expansion during the 2000s. However, the short run elasticity with respect to output is rather low (according to the static model), ranging between 0.31 (Hungary) and 0.57 (Czech Republic); the short-run elasticity becomes even smaller in the dynamic estimations, with the short-run elasticity ranging between 0.10 in Slovenia and 0.35 in Slovakia, since the lagged effect of employment is also controlled for. However, in these cases the long-run elasticities reach relatively high levels (0.70 in Slovakia, 0.75 in Slovenia, and 1.46 in Bulgaria, the latter of which is too high, but not robust in different sector groups as will be discussed below), but the speed of adjustment is relatively low. Comparing these elasticities with those estimated by Basu et al. (2005) for the transition period (the periods differ between 1990 to 1993 for different countries), both the short and long-run elasticity for Poland and Slovakia are quite close to the estimations here for the period after 1999; our estimates for Hungary are lower but not too different; but for the Czech Republic we find significantly lower elasticities. The difference can be due to
86
Özlem Onaran
the initial high response of employment in the Czech Republic to the big bang recession with a dramatic decline in labor hoarding, and once the over employment and recession is over, firms might have become less responsive to expansion under competitive pressures. In the pool of HS industries, the coefficient of output is in the majority positively significant (all but Lithuania, where the coefficient is even negatively significant) compared to the MLS sectors (significant only in five countries: Hungary, Poland, Slovenia, Slovakia, Lithuania). However, the demand elasticity of employment is also lower in the HS sectors compared to the MLS sectors in five countries (Hungary, Poland, Slovenia, Slovakia, Lithuania). Thus in these countries, labor saving growth is stronger in the HS sectors compared to the MLS sectors in the short-run. However when the long-run elasticities are compared across sector groups, in all the three cases where lagged employment is significant in both sector groups (Slovakia, Bulgaria, Romania), the long-run elasticity is higher in the HS sectors. So although labor saving may dominate in the HS sectors in the short-run, in these three countries the longrun effects of growth in skilled industries may be more favorable than in low-skill sectors. However, it must be also added that the long-run elasticities in the sector groups are much lower than those calculated based on the total pool: in the HS sectors the long-run elasticity is 0.27 in Slovenia, 0.50 in Romania, 0.53 in Bulgaria, and 0.92 in Slovakia, which is the only country with a high elasticity. The case of Bulgaria is particularly interesting since the long-run elasticity for the pool of all sectors was even greater than one. In the MLS sectors the long-run elasticities are as follows (only for the countries, where the dynamic specification is valid): 0.30 in Bulgaria, 0.89 in Slovakia, 0.99 in Czech Republic. The effect of real wage on labor demand is mixed across countries. According to the results for the pool of all sectors, the classical assumption of a negative effect of wages on labor demand is only valid in three countries (Slovakia, Romania Lithuania). In all other countries the effect is insignificant. In the HS sectors, in four countries (Hungary, Slovenia, Romania, Czech Republic) wages have a negative significant effect on employment. In three countries (Slovakia, Lithuania, Poland) wages are again insignificant in the HS sectors. In Bulgaria the effect is even positively significant. This controversial finding for Bulgaria is also the case in the MLS sectors, consistent with the stylized facts indicating a par-
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
87
allel decline in both employment and real wages in the MLS sectors. For Poland also a positive wage coefficient is found. The effect of wages is insignificant in the MLS sectors in five countries (Hungary, Slovenia, Czech Republic, Slovakia, Romania). Only in Lithuania a negative significant wage effect can be found in the MLS sectors. Thus the job performance of most economies, particularly in the MLS sectors, can not be explained by the developments in wages. In the HS sectors labor demand is sensitive to wages in half of the countries, but again in the other half low wages do not lead to higher employment. Finally, if we compare the absolute value of the effect of output and wages on employment in the three cases (pool of all sectors, HS and MLS sectors) in total manufacturing, only in Slovakia the output effect is relatively larger. In the HS sectors in all four countries, and in the MS sectors in Lithuania, which is the only country with a significant negative wage effect, the wage effect is dominant. So in the minority cases where there is a negative and significant wage effect, wage moderation seems to be a relatively more effective strategy to create jobs compared to demand policies. In the rest of the cases wage competition does not cure the problem of low employment. Regarding the effect of foreign trade, exports to EU15 and imports from EU15 as a ratio to output are added to the basic specification as explanatory variables. The results for foreign trade are robust when total exports and imports are used instead of those to/from EU. According to the results for the pool of all sectors, export ratio has a robust positive effect on employment only in Romania. In all other countries (Hungary, Poland, Slovenia, Slovakia, Czech Republic, Bulgaria, Lithuania) the effect is insignificant. The positive effects of exports on employment in Romania is limited only to the HS sectors, with a 1 percentage point increase in the export ratio leading to a 0.53 per cent increase in employment in the HS sectors (in the MSL sectors the effect is insignificant). In the HS sectors, the effect is also positive in Lithuania, with a similar degree of economic significance as in Romania. In the HS sectors the effect of exports are negatively significant in Slovakia and robust to specifications. A 1 percentage point increase in the export/output ratio (to EU15) in the HS sectors in Slovakia is leading to a 0.33 per cent decrease in employment. In five countries exports have no significant effect on employment in the HS sectors. In the MLS sectors, there is no country with a positive export effect
88
Özlem Onaran
on employment. To summarize, in five countries (Hungary, Poland, Slovenia, Czech Republic, Bulgaria) the employment effect of exports is insignificant in both HS and MLS sectors. In Slovakia in the HS sectors exports have a negative effect. Only in Lithuania and Romania and only in the HS sectors there is a positive effect of exports. The effect of the import ratio is also insignificant in the majority of the countries. Only in Bulgaria there is evidence for a positive significant effect according to the results of the total pool, however this effect cannot be robustly verified at the sectoral levels. In HS sectors the effect of imports is significantly negative in Lithuania and Slovenia. However, this negative effect is economically not very significant: a 1 percentage point increase in imports from EU/output in Slovenia decreases HS employment by 0.07 per cent. The effect is even lower in Lithuania. The effect of imports is positive in Hungary in the MLS sectors, and the economic significance is also relevant with a 1 per cent-point increase in imports from EU/output in the MLS sectors leading to a 0.43 per cent increase in employment. Other than the HS sectors in Slovenia and Lithuania, there is no evidence of a negative demand or efficiency induced effect of imports on employment. In Hungary the positive effect of imports on labor demand in the MLS sectors can be an indicator that imports are complements rather than substitutes for domestic production. Finally, FDI stock as a ratio to output is added as an explanatory variable in addition to output and real wages. According to the total pool results, FDI has a significant positive effect on employment only in Lithuania, but the effect cannot be verified in the sector groups for the HS or MLS sectors. In all other five countries the effect is insignificant. Bulgaria and Romania are not included due to lack of FDI data. FDI has a positive effect on employment at a sectoral level only in the MLS sectors in Slovakia. But in the HS sectors in Slovakia and Slovenia the effect of FDI on sectoral employment is negative. These results are robust when FDI is used without the output variable. Hunya and Geishecker (2005) also estimate the effect of FDI on the skill composition of employment in the CEECs, and find that FDI results in more employment of highskilled non-manual workers and low-killed workers, although the magnitude of this effect is modest. To compare their results with those of this paper, if we assume that manual workers are dominating the trends in total manufacturing employment in the HS sectors,
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
89
the findings of Hunya and Geishecker is in line with the findings of this paper for the negative effects in the HS sectors in Slovakia and Slovenia, and positive effects in the MLS sectors in Slovakia. However, this is not a general finding of this paper, since the effects are insignificant in all other cases. V. Conclusion
The paper has analyzed the links between employment and domestic and international factors in the CEECs based on the case of manufacturing industry during the period after the transition recession. Regarding the domestic factors, an interesting finding is that employment does not respond to wages in more than half of the cases, thus jobless growth in manufacturing is taking place irrespective of the wage developments in the majority of the cases. The response of employment to output is mostly positive, however the output elasticity of labor demand is rather low in the short-run, and in four cases lower in the high-skill industries. There are also a number of cases where employment is completely de-linked from output, mostly in the medium and low-skill sectors (in the Czech Republic, Bulgaria and Romania). Although the starting conditions before transition indicate underutilization of labor and therefore high employment rates in these countries, the process of the transition shock reversed these conditions significantly; therefore further downsizing almost a decade after transition indicate the relevance of international competitive pressures in determining the continuation of downsizing and productivity enhancing job destruction in the 2000s. Wherever partial adjustment model was significant, the long-run output elasticities were also calculated, and although they were higher than the short-run elasticities, in many cases, particularly at the level of sector groups, the values were still not very high, and the speed of adjustment is found to be low. Regarding international factors, it can be concluded that overall an impressive speed of integration into the European economic sphere through FDI and international trade has not prevented job losses in the manufacturing industry. Among eight CEECs, only in Romania and Lithuania and only in the high-skill sectors exports have the expected positive effect on employment. In Slovakia the effect of exports on employment in the high-skill sectors has been negative. The insignificant effects in the MLS sectors are particularly inconsistent with the optimistic expectations regarding increased labor demand in sectors where these countries are
90
Özlem Onaran
supposed to have their comparative advantage. The positive effect of exports in the high-skill sectors is also a result not expected according to traditional trade theory, but is consistent with the arguments suggesting an increased skill bias of foreign trade. The negative effect of exports in the high-skill sectors in Slovakia is in line with both the comparative advantage based explanations of traditional trade theory and the arguments about the efficiency inducing effects of foreign trade dominating job creation effects. Regarding imports, in the high-skill sectors a negative effect on employment is found in Slovenia and in Lithuania. In Hungary, in the medium and low-skill sectors imports have a positive impact on employment. This positive effect indicates the complementary character of imports rather than being a substitute for domestic production – a result related to the import dependency of production. Finally, FDI has a significantly positive effect on employment only in the medium and low-skill sectors in Slovakia, but a negative effect on employment in the high-skill sectors in Slovakia and Slovenia, where the induced efficiency effects of FDI and the possible negative spillover effects on domestic firms are dominating the positive demand effects. There is no significant robust effect in the other countries. The results suggest that the positive demand effects of the integration of the CEECs into the world economy have been offset by international competitive pressures, leading to labor saving growth without generating jobs. These results combined with the flexibility in the institutional structure of the labor market and the wage setting mechanism, show that the source of the employment problem results from the demand side of the labor market, rather than the supply side. Until now the catching-up and integration process, which involved specialization in export-oriented manufacturing sectors, where the CEECs have a low-wage advantage compared to Western Europe, also necessitated a continuous productivity improvement based on declining employment in these sectors, since sectoral productivity levels were nevertheless lower than those in the West. However, this is also leading to a deviation between two outcomes: the success in terms of growth led by exports and FDI vs. the stagnant high unemployment or low employment rates. This divergence in turn can lead to a deterioration in the popular support for European integration and create the ground for political instability.
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
91
This pattern seems hard to reverse as long as the economic policy relies mostly on the highly crowded export markets as a source of demand, and low wage increases and labor saving as a key to resist competitive pressures, which in turn leads to a bottleneck regarding domestic demand. However, the small size of the domestic markets in these countries also does not leave much area for maneuver to reverse this pressure. This dilemma can only be overcome via an EU level coordinated macroeconomic and industrial policy targeting employment increase along with wage and productivity convergence in the medium and the long run. Indeed, as Western European workers fear that relocation and outsourcing to the East lead to job losses in the West, the same process is not necessarily leading to job creation in the East. This perverse situation prepares the ground for cooperation and coordination between the East and the West. Regarding the FDI related aspects, industrial policy to promote not only greenfield investments and reinvestment of profits but also strong backward linkages with the domestic economy may facilitate job creation via positive spillover effects.
References
Amsden, Alice H., Hoeven, Rolph Van Der (1996), Manufacturing output, employment and real wages in the 1980s: labor’s loss until the century’s end, The Journal of Development Studies 32 (4), 506-530. Basu, Swati, Estrin, Saul, Svejnar, Jan (2005), Employment Determination in Enterprises under Communism and in Transition: Evidence from Central Europe, Industrial and Labor Relations Review 58 (3), 353-69. Blanchard, Olivier J. (1998), The economics of Postcommunist Transition, Claredon Press, Oxford. Boeri, Tito, Garibaldi, Pietro (2006), Are Labour Markets in the New Member States Sufficiently Flexible for EMU?, Journal of Banking and Finance 30 (5), 1393-1407.
92
Özlem Onaran
Boeri, Tito, Terrell, Katherine (2002), Institutional Determinants of Labor Reallocation in Transition, Journal of Economic Perspectives 16 (1), 51-76. Brada, Josef C. (1989), Technological Progress and Factor Utilization in Eastern European Economic Growth, Economica 56 (224), 433-48. Feenstra, Robert C., Hanson, Gordon H. (1996), Foreign direct investment, outsourcing, and relative wages, National Bureau of Economic Research Working Paper no. 5121. Gora, Marek (1995), The Labor Market in Poland: A Tentative Analysis of the First Years of Transition, Eastern European Economics 33 (5), 75-96. Ghosh, Jayati (2005), Exporting Jobs or Watching them Disappear? Relocation, Employment and Accumulation in the World Economy, in: Ghosh J. and Chandrashekar C. P. (eds.), Work and Well-Being in the Age of Finance, New Delhi, Tulika Books, 99 – 119. Greenaway, David, Hine, Robert C., Wright, Peter (1999), An empirical assessment of the impact of trade on employment in the United Kingdom, European Journal of Political Economy 15, 485-500. Havlik, Peter, Landesmann, Michael (2005), Structural change, productivity and employment in the new EU member states, in: Economic restructuring and labour markets in the accession countries, The Vienna Institute for International Economic Studies Research Project commissioned by EU DG Employment, Social Affairs and Equal Opportunities. Hine, Robert C., Wright, Peter (1997), Trade and manufacturing employment in UK manufacturing, in: Borkakoti, Jitendralal, Milner, Chris (eds.), International Trade and Labour Markets, London, Macmillan, 118-139. Hunya, Gábor, Geishecker, Ingo (2005), Employment effects of foreign direct investment in Central and Eastern Europe, The Vienna Institute for International Economic Studies Research Reports, no. 321. Horton, Susan, Kanbur Ravi, Mazumdar Dipak (1994), Labor markets in an era of adjustment: an overview, in: Horton, Susan,
The Effect of Trade and FDI on Employment in Central and Eastern European Countries
93
Kanbur, Ravi, Mazumdar, Dipak (eds.), Labor Markets in an Era of Adjustment 1, Washington, D.C, World Bank, 1-60. Izyumov, Alexei, Vahaly, John (2002), The Unemployment-Output Tradeoff in Transition Economies: Does Okun's Law Apply?, Economics of Planning 35 (4), 317-31. Jackman, Richard (1994), Economic Policy and Employment in the Transition Economies of Central and Eastern Europe: What Have We Learned?, International Labour Review 133 (3), 32745. Kornai, Janos (1995), Highways and Byways: Studies on Reform and Postcommunist Transition, Cambridge, MA: MIT Press. Kajzer, Alenka (1995), The Real-Wage-Employment Relationship and Unemployment in Transition Economies: The Case of Slovenia and Hungary, Eastern European Economics 33 (4), 5578. Landesmann, Michael, Vidovic, Hermine, Ward, Terry (2004), Economic restructuring and labor market developments in the new EU member states, The Vienna Institute for International Economic Studies Research Report, no. 312. Lehmann, Hartmut (1995), Active Labor Market Policies in the OECD and in Selected Transition Economies, The World Bank, Policy Research Working Paper Series no.1502. Mencinger, Joze (2003), Does foreign direct investment always enhance economic growth?, Kyklos 56 (4), 491-508. Milner, Chris, Wright, Peter (1998), Modelling labour market adjustment to trade liberalisation in an industrialising economy, The Economic Journal 108, 509-5028. OECD (2004), Employment Outlook, Paris. Onaran, Özlem (2001), The effect of trade liberalisation on labor demand in Turkish manufacturing industry, Annual Conference of the Association of Heterodox Economics, London, July 7-8 2001. Onaran, Özlem (2008), Jobless growth in Central and Eastern European Countries: A country-spcific panel data analysis for the manufacturing industry, Eastern European Economics 46 (4), 90-115.
94
Özlem Onaran
Onaran, Özlem, Stockhammer, Engelbert (2008), The effect of FDI and foreign trade on wage bargaining in the Central and Eastern European Countries in the post-transition era: A sectoral analysis, Structural Change and Economic Dynamics 19, 66-80. Pollin, Robert, Burns, Justine, Heintz, James, (2004), Global apparel production and sweatshop labour: can raising retail prices finance living wages?, Cambridge Journal of Economics 28, 153-171. Rodrik, Dani (1997), Has globalization gone too far?, Institute for International Economics, Washington DC. Reinert, Erik S., Kattel, Rainer, (2004), The Qualitative Shift in European Integration: Towards Permanent Wage Pressures and a ‘Latin-Americanization’ of Europe?, PRAXIS Center for Policy Studies Working Paper no. 17. Schaffer, Mark E. (1998), Do Firms in Transition Economies Have Soft Budget Constraints? A Reconsideration of Concepts and Evidence, Journal of Comparative Economics 26, 80–103. Stehrer, Robert (2004), Can trade explain the sector bias of skillbiased technical change?, Vienna Institute for International Economic Studies Working paper no. 30. Wood, Adrian (1994), North-South Trade, Employment and Inequality: Changing Fortunes in a Skill-Driven World, Oxford: Clarendon Press. World Bank (2006), Doing Business Report, Washinton D.C.
Svetlana Ledyaeva
Foreign Direct Investment Determinants in PostCrisis Russia: Empirical Evidence1 Abstract In this paper, we provide an empirical analysis of the determinants of FDI inflows into Russian regions for the post-crisis period (1999-2005). Using cross-section OLS we obtained the following results. First, the main determinants of FDI inflows into Russian regions during transition were market size, the presence of big cities, oil and gas resources and political and legislative risk factors. The hydrocarbon-endowed Sakhalin region almost stands as an FDI determinant by itself. I. Introduction During the last 15 years – the period while the Russian economy has been transforming from a command to market economy and as a consequence liberalizing its international economic relationships including openness to foreign direct investments (FDI) – Russia` s achievements in attracting FDI have been quite modest. The share of Russia in the world FDI stock in 2002 was 0.3 per cent, while, for example China had received 6.3 per cent of the world` s total FDI stock. If we look at the per capita FDI stock in Russia, the picture is even worse. Russia has received 24 times less per capita FDI than the Czech Republic and 12 times less than countries such as Estonia, Slovenia and Hungary (UNCTAD, 2003). In a survey among foreign investors in Russia, most of the problems mentioned as barriers to investment were of an institutional and legislative nature (Liuhto et al. 2004). Besides the small amounts of inward FDI into Russia, the industrial and regional structure of attracted FDI seems to be inconsistent and ineffective in the context of industrialization and postindustrial development of the economy. This concerns also the 1
An extended version of this paper is forthcoming under the title ”Spatial Econometric Analysis of Foreign Direct Investment Determinants in Russian Regions” in the World Economy Journal.
96
Svetlana Ledyaeva
more or less unequal regional development in Russia. The uneven distribution of FDI can partly be explained with Russia’s large territory but this does not alone explain the distribution problem and the low scale of FDI’ s in Russia in general. Our research aims to determine the main factors determining inward FDI into Russia’s regions. We estimate our specification in order to analyze the main determinants of FDI in Russia’s regions in the period of 1999-2000. We focus on the post-crisis period as FDI inflows increased considerably in this period in comparison with the period before the financial crisis of 1998. The structure of paper is the following. Section II describes the regional distribution of FDI in Russia. In the last decades a large amount of literature has developed on empirical evidence of FDI flows. In section III we briefly discuss the existing empirical literature on FDI determinants in Russia and other transition economies. Section IV describes the methodology and data of the study. Section V summarizes the results of empirical estimations, and section VI concludes the paper. II. Regional distribution of FDI across Russia The aggregate development of FDI flows into the Russian economy in the period of 1996-2005 is presented in Figure 1. Before 2002, aggregate FDI inflows into Russia were low, but stable, with a small peak in 1997. Since 2003, a strong increase of FDI inflows into Russia can be observed. Figure 1: Development of FDI inflows in Russia, 1996-2005 14000 million US-$
12000 10000 8000 6000 4000 2000 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Source: Rosstat statistics.
Foreign Direct Investment Determinants in Post-Crisis Russia
97
Our goal here is to shed light on the FDI determinants across Russian regions in the post-crisis period of 1999-2005. To gain first evidence about the FDI distribution across Russian regions2 and their changes throughout the analyzed period, we report Russia’s top ten recipient regions of FDI. These “top 10” regions have received approximately 70-80 per cent of total FDI inflows into Russia. Thus, our analysis of the top 10 regional recipients of FDI should be quite representative when determining the crucial FDI determinants for Russia. The top 10 regional recipients of FDI in post-crisis period are presented in Table 1. Table 1: Top 10 Russian FDI recipient regions in the post-crisis period (1999-2005) top 10 regions absolute top 10 regions per capita FDI share top 10 regions FDI inflows FDI region cum. FDI, region avg FDI region avg. FDI/ million p.c., regional dollars, dollars, GDP (%), 1999199919992005 2005 2005 Sahalin 13867.3 Sahalin 3467 Sahalin 128.4 Omskaya Omskaya oblast Moscow 4288.5 oblast 187 16.4 Omskaya Lipeckaya Novgorodskaya 2772.2 oblast 121 oblast 7.7 oblast Moscovskaya Magadanskaya Lipeckaya oblast 1998.7 oblast 105 oblast 7.5 Novgorodskay Kostromskaya Tjumenskaya oblast 1574.4 a oblast 98 oblast 6.2 Lipeckaya Leningradskay Leningradskaya oblast 1034.5 a oblast 74 oblast 4.5 Leningrad Tjumenskaya Magadanskaya oblast 861.5 oblast 69 oblast 4.1 Krasnodarskij Moscow Kaluzhskaya kraj 792.4 66 oblast 3.1 Novgorodska Kostromskaya Moskovskaya ya oblast 486.0 oblast 62 oblast 3.0 SaintMoskovskaya Vladimirskaya Petersburg 470.8 oblast 44 oblast 2.3
2
There are 89 Russian regions. However we use only 74 regions due to data constraints.
98
Svetlana Ledyaeva
From Table 1 we conclude that FDI is concentrated in natural resources` abundant regions (Sahalin, Tjumenskaya oblast, Magadanskaya oblast), the two biggest cities, Moscow and Saint Petersburg, and their territories and also some other regions which are mostly located in central Russia (Vladimirskaya oblast, Kaluzhskaya oblast, Lipetckaya oblast and Novgorodskaya oblast). A remarkable feature is that the oil and gas abundant Sakhalin region is leader with respect to all reported FDI indicators. We also use a map to illustrate the geographical distribution of cumulative FDI in the period (see Figure 2). Figure 2: Regional distribution of cumulative FDI across Russian regions, 1999-2005 (US$ million)
Top 10: 470-13870 millions of US$ 235-470 millions of US$ 94-235 millions of US$ 47-94 millions of US$ Less than 47 millions of US$ Lakes Data is not available white Source: Author’s own calculations based on Rosstat statistics
In general, we observe that there are no evident geographical patterns in FDI distribution across Russia in both periods. In any
Foreign Direct Investment Determinants in Post-Crisis Russia
99
geographical part of Russia (West, East, North and South), there are regions with different levels of FDI. III. Previous empirical studies on FDI determinants in Russia and other transition/emerging economies A. Empirical research on FDI determinants in Russia To the best of our knowledge there are only four empirical studies which focus on the determinants/factors of FDI inflows into Russia during transition based on regional data. In our paper we attempt to contribute to the existing literature using a rather new methodology and trying to avoid problems arisen in previous studies. Brock (1998) analyzes FDI determinants in the early transition – 1993-1995 years. The author found that market size and the crime situation are important for FDI. Another interesting result is that education of labour force is important only for FDI into Moscow and Saint Petersburg. Broadman and Recanatini (2001) analyse determinants of FDI inflows in the period 1995 to 1999 using GLS estimation for panel data and OLS estimation for cross-section data. They found that market size, infrastructure's development, and policy framework factors explain much of the observed variation of FDI inflows across Russia's regions. Their results also suggest that the model that explains well the cross-regional variation in FDI flows from 1995-1998 changes significantly in terms of explanatory power following the 1998 crisis, suggesting a "structural regime change" in the FDI framework in Russia in the post-crisis period. Iwasaki and Suganuma (2005) suggested a model of the regional distribution of FDI in Russia based on panel and yearly crosssectional data for 1996–2003. The authors conclude that resource endowments, market factors and industrialization and infrastructural factors proved to have high significance and explanatory power in the empirical analysis. In addition, they suggested that climate and measures of regional governments to promote FDI are possible investment factors. No evidence was found in that analysis that the Russian financial crisis in 1998 had a statistically significant influence on the decision-making process of foreign investors. Ledyaeva and Linden (2006) estimate a gravity model of inward FDI to determine the sources of uneven distribution of FDI across Russian regions in recent years. The authors use OLS and binary dependent variable models to analyse the determinants of the
100
Svetlana Ledyaeva
number of foreign firms registered until the year of 2002. The OLS results suggest that gross products of host regions and source countries, agglomeration effects, Moscow city advantages, cultural closeness and skilled labour abundance are positively related to the number of foreign firms in a particular Russian region. The distance between host regions and source countries is negatively related to the dependent variable. As for binary choice analysis the results show that only four factors can be considered to be important in determining the probability of a foreign firm entering in a particular Russian region. These factors are gross product of host regions and source countries, distance between them and agglomeration effects. In general, previous studies emphasize the importance of such FDI determinants into the Russian economy as market size, resource abundance, industrialization and infrastructural factors, policy framework factors and skilled labour force abundance. B. FDI determinants in transition/emerging economies In this section we review the most recent empirical studies on FDI determinants in transition/emerging economies. Globerman, Shapiro and Tang (2006) examine the determinants of both inward and outward FDI for twenty emerging and transition economies over the period 1995-2001. As inward FDI determinants they analyze GDP, GDP growth, a governance index, the ratio of imports to exports, stock market capitalization, privatization, an oil indicator, dummy variables for China, EU members, future EU members, euro currency and regional dummy variables. They found that market size, the governance index and privatization are positively related and that the oil indicator is negatively related to FDI inflows. Another interesting result is that joining the EU, or even the prospect of joining the EU, promotes inward FDI (a halo effect), and this phenomenon particularly characterizes the former Communist countries. Janicki and Wunnava (2004) examined bilateral foreign direct investments (FDI) between the members of the European Union and eight Central and Eastern European countries (CEECs) in transition, awaiting accession into the European Union (EU). This study reveals that the key determinants of FDI inflows in the CEECs are the size of the host economy, host country risk, labour costs in the host country, and openness to trade. Frenkel, Funke and Stadtmann (2004) examine the determinants of FDI flows to emerging economies by analyzing a recently
Foreign Direct Investment Determinants in Post-Crisis Russia
101
compiled data set of bilateral FDI flows. Using a panel approach, they investigate both home and host country factors that may play an important role in determining the level and the destination of FDI flows. On the home country side of the FDI flows, they focus on the five largest economies worldwide. On the host country side, they study a number of emerging economies in Asia, Latin America, and Central and Eastern Europe. They find that the gravity approach can indeed successfully be applied to FDI. While market size and distance, which are classical explanatory variables in the gravity model, play an important role for FDI flows, other economic characteristics brought forward in FDI theory like, e.g., risk and economic growth in host countries are also crucial for attracting international investment projects. Finally, growth in countries from which FDI activities originate, exerts a positive effect on the level of FDI flows so that such activities also depend on the business cycles in the home countries. Carstensen and Toubal (2004) use dynamic panel data methods to examine the determinants of foreign direct investment (FDI) in CEECs. Their empirical model shows that the traditional determinants, such as market potential, low relative unit labour costs, a skilled workforce and relative endowments, have significant and plausible effects. In addition, transition-specific factors, such as the level and method of privatization and the country risk, play an important role in determining the flows of FDI into the CEECs and help to explain the differing attractiveness of the individual countries to foreign investors. Bevan and Estrin (2004) use a panel dataset of bilateral flows of foreign direct investment (FDI) and study the determinants of FDI from Western countries, mainly in the European Union (EU), to Central and Eastern European ones. They find the most important influences to be unit labour costs, gravity factors, market size, and proximity. Interestingly, host country risk proves not to be a significant determinant. Their empirical work also indicates that announcements about EU Accession have an impact on FDI for the future member countries. We can summarise that market size, country's risk factors, labour costs, methods and level of privatization and future prospects of EU accession are crucial factors of FDI inflows into transition/emerging economies.
102
Svetlana Ledyaeva
Comparing empirical evidence on FDI determinants in Russia and other transition/emerging economies we conclude that findings are similar in general. However there are some differences in the results. This gives a motivation for further empirical investigation of FDI determinants in the Russian economy as a somehow unique case among transition economies due to its huge territory and former status of super empire in the socialist block. IV. Model and data description We start our model discussion with a look at earlier studies based on yearly FDI flows. The explanatory power of the studies of Broadman and Recanatini (2001) and Iwasaki and Suganuma (2005) suffer from the impacts of extraordinary years or projects that boost a region’s gross FDI.3 Thus, we take as our dependent variable the natural logarithm of cumulative FDI inflows in the period of 2000-2005. All explanatory variables are taken in original terms in the year 1999. The exception is the market size variable, which is taken as the average for the period one-year lagged from the dependent variable’s period (1999-2004) as constructed to give it greater explanatory power than market size one-year lagged from the cumulative period of the dependent variable. This might indicate flexibility in market-seeking FDI to short-term changes in market size. The use of lagged explanatory variables helps to solve possible endogeneity problems and relates to a simple hypothesis for the foreign investor decision-making. Foreign investors are assumed to make an investment decision for a given year/period by referring to the observable variables of the previous year (e.g. Coughlin and Segev, 2000; Iwasaki and Suganuma, 2005). Through trial and error, seven FDI determinants are selected. According to the survey results obtained in the previous section, these variables relate to the following five factors: market size, infrastructure development, policy framework, resource endowments and industrialization. The data sources of the explanatory variables are summarized in Appendix 1. In particular, we specify:
a0 a1msi a2 porti a3bci + + a4 dsahi a5oil _ga _ indexi a6leg i a7 poli H i ln FDI i
3
See also Xing and Kolstad (1997) and Brock (1998).
(1)
Foreign Direct Investment Determinants in Post-Crisis Russia
103
The market size variable msi is composed of three variables (gross regional product, total population and population density in a region i). The same indicator for market size is used in the study of Iwasaki and Suganuma (2005). Results from principal component analysis of market size are represented in Appendix 2. The proportion of variance of the first component reaches 80 per cent, and furthermore, its eigenvector and component loading show that this variable is suitable as a general index of the market size. The variable is taken as average for the period one-year lagged from the dependent variable’s period. The next variable, porti , is the number of ports in region i. It proxies transport infrastructure for a particular region. bci is a dummy variable for regions that include at least one of Russia’s 13 cities with populations exceeding 1 million.4 This is a proxy for the level of industrialization in a particular region. dsah is a dummy variable for the Sakhalin region, where large product sharing agreements in oil and gas industries have been launched during transition. The variable oil _ gas _ indexiis constructed on the basis of two variables for oil and gas production in region i (for details see Appendix 3). We use the formula of integrated coefficient adopted from Ndikumana (2000). leg i is the legislative risk in a region i and poli is a political risk in region i. We use the calculation for these risks published by the Russian economic journal Ekspert. an are parameter estimates and H i is the error term. In our preliminary estimations we also considered other explanatory variables (climate, average wage in a region, etc.), however the test of omitted variables rejected the inclusion of these variables into our final specification. To estimate specification (1) we use cross-section OLS. V. Results
Table 2 presents cross section OLS results of Equation (1) for the entire sample.
4
The 13 cities are Moscow, St. Petersburg, Novosibirsk, NizhnyNovgorod, Ekaterinburg, Samara, Omsk, Kazan, Chelyabinsk, Rostovna-Donu, Ufa, Volgograd, Perm.
104
Svetlana Ledyaeva
Table 2: Estimation results, cross-section OLS results, dependent variable: cumulative FDI in region i in 2000-2005. variable Constant
msi porti
coefficient 4.4 (9.5)*** 0.28 (2.6)** 0.13 (1) 1.24 (2.5)**
bci dsah oil _gas _ indexi leg i poli
-0.0002(-0.03)
No. obs. Adjusted R-square Normality test Heteroskedasticity test
61 0.44 2.7 (0.27) 20.9 (0.79)
4.8 (3)*** 0.05 (2.2)** -0.016 (-2.2)**
Note: t-statistics in parentheses; ***, **, * denote the 1, 5 and 10 per cent significance level. Normality test based on a chi-square distribution, null hypothesis: normal distribution of errors. White test for heteroskedasticity, null hypothesis: homoskedasticity of errors.
From the results we conclude that the important determinants of FDI inflows into Russian regions during the post-crisis period have been market size, the presence of big cities in a region, oil and gas resources, legislative risk. The Sakhalin region experienced overproportionate FDI inflows. Sea ports and political risk in a particular Russian region do not show significant relationship with FDI regional distribution. VI. Conclusions and policy implications
In this paper, we empirically analyzed the determinants of FDI inflows into Russian regions during the post-crisis period (19992005). Using cross-section OLS we obtained the following results. First, the major determinants of FDI inflows into Russian regions during transition have been market size, the presence of big cities, oil and gas resources and political and legislative risks. The hydrocarbon-endowed Sakhalin region almost stands as an FDI determinant unto itself. The following policy implications are coming from our study. First, regional legislative risk has become an important factor of
Foreign Direct Investment Determinants in Post-Crisis Russia
105
FDI inflows into Russia in recent years. This means regional authorities must pay increased attention as to how legislative initiatives are crafted if they wish to attract FDI. Second, market size and the presence of large cities are important factors for FDI inflows into Russia, especially in recent years. Thus, managed growth and urbanization will assure that cities remain attractive destinations for foreign investors. Indeed, FDI can help them in emerging as leaders of industrial and economic development in Russia. Therefore, federal and regional authorities need to take measures that protect and promote favourable investment climates. Moreover proximal regions can also benefit from FDI, so improving transport connections between urban centres and nearby regions, along with improving the general investment climate in nearby regions, should also be noted.
References
Bevan, Alan A., Estrin, Saul (2004), The determinants of foreign direct investment into European transition economies, Journal of Comparative Economics 32 (4), 775-787. Broadman, Harry G., Recanatini, Francesca (2001), Where Has All the Foreign Investment Gone in Russia?, Policy Research Working Paper No. 2640, Washington, DC, World Bank. Brock, Gregory J. (1998), Foreign Direct Investment in Russia’s Regions 1993–95, Why So Little, and Where Has It Gone?, Economics of Transition 6 (2), 349-360. Carstensen, Kai, Toubal, Farid (2004), Foreign direct investment in Central and Eastern European countries: a dynamic panel analysis, Journal of Comparative Economics 32 (1), 3-22. Coughlin, Cletus, Segev, Eran (2000), Foreign direct investment in China: a spatial econometric study, The World Economy 23(1), 1-23. Frenkel, Michael, Funke, Katja, Stadtmann, Georg (2004), A panel analysis of bilateral FDI flows to emerging economies, Economic Systems 28 (3), 281-300.
106
Svetlana Ledyaeva
Globerman, Steven, Shapiro, Daniel, Yao, Tang (2006), Foreign direct investment in emerging and transition European countries, International Finance Review 6, 431-459. Iwasaki, Ichiro, Suganuma, Keiko (2005), Regional Distribution of Foreign Direct Investment in Russia, Post-Communist Economies 17 (2), 153-172. Janicki, Hubert P., Wunnava, Phanindra V. (2004), Determinants of foreign direct investment: empirical evidence from EU accession candidates, Applied Economics 36 (5), 505-509. Ledyaeva, Svetlana, Linden, Mikael (2006), Testing for Foreign Direct Investment gravity model for Russian regions, Department of Business and Economics Working paper No.32, University of Joensuu. Liuhto, Kari - Pelto, Elina - Lipponen, Kirsi (2004), Where to Do Business in Russia? A Report on Russian Regions, Firms, Foreign Trade and Investment Flows, Pan-European Insitute, Turku, Finland (http://www.tse.fi/FI/yksikot/ erillislaitokset/pei/Documents/Julkaisut/Liuhto_Pelto_Lipponen _52004.pdf) Ndikumana, Leonce (2000), Financial determinants of domestic investment in Sub-Saharan Africa: Evidence from panel data, World Development 28 (2), 381-400. Unctad (2003), World Investment Report, Unctad, Geneva. Xing, Yuquing, Kolstad, Charles (1997), Do Lax Environmental Regulations Attract Foreign Investment?, Working Paper in Economics, 6-95R, UC Santa Barbara. Xing, Yuquing, Kolstad, Charles (2002), Do Lax Environmental Regulations Attract Foreign Investment?, Environmental & Resource Economics, European Association of Environmental and Resource Economists 21, 1-22.
Foreign Direct Investment Determinants in Post-Crisis Russia
107
Appendix 1 Table A1.1: Explanatory variables used in the specification var.
definition data source Rosstat Market size –first msi principal component of three variables – GRP, total regional population and regional pop. density (for calculateon details, see App. 2, Table A2.1) http://www.transrussia.net/ports/port.aspx Number of sea porti ports in a region Dummy variable of bci 13 biggest cities in Russia Equal to 1 for Sakhalin region, 0 dsah Dummy for otherwise Sakhalin region oil _ gas _ Oil and gas index Rosstat (for calculation indexi details see App. 2) Legislation risk Russian economics journal Ekspert legi (http://www.raexpert.ru/ratings/regions/) Political risk Russian economics journal Ekspert pol i
Appendix 2 Table A2.1: Results from principal component analysis of market size Principal components/correlation Component Eigenvalue Difference Proportion Cumulative Comp 1 2.39692 2.02709 0.7990 0.7990 Comp 2 .369832 .136588 0.1233 0.9223 Comp 3 .233244 0.0777 1.0000 Principal components (eigenvectors) Variable Comp1 Comp2 Comp3 Unexplained GRP 0.5950 -0.0739 -0.8003 0 Pop 0.5650 0.7467 0.3511 0 Pop_density 0.5717 -0.6611 0.4860 0 Note: PCA is based on panel data set for 74 Russian regions over the period 1999-2004. The resulting panel vector of first principal component was used to construct final market size variables as averages of the period.
108
Svetlana Ledyaeva
Appendix 3
The oil and gas index was calculated using the following formula of integrated coefficient: Oil _ gas _ indexit
1
m
ª
j 1
¬«
§ Fj ,it · º ¸¸ », © jt ¹ ¼»
¦ «100 * ¨¨ F m
where i=1,…,74 in period t=1996,…, 2004. Fj ,it is the actual resource indicator j for a region i in period t, Fjt is the sample mean of the indicator in period t (in our case the mean value for Russian 1 n Fijt , where n is the number of Russian regions, which is Fjt n i1 regions involved in the computation(74)), m is the number of indicators included in the index computation (adopted from Ndikumana, 2000). Indicators, included in the computation of the resource index are presented in Table A1.1.
¦
Table A3.1: Indicators included in the Resource Index N 1 2
Indicator Oil digging including gas condensate per capita, thousands of tones Natural gas digging per capita, millions cubic meters
Part 3: The EU and Latin America
111
Inmaculada Martínez-Zarzoso*
On Transport Costs and Sectoral Trade: Further Evidence for Latin-American Imports from the European Union Abstract In this paper we evaluate the relative importance of different sources of trade costs, namely, freight rates, poor infrastructure, and location with respect to Latin-American imports from the European Union (EU). Firstly, a transport-costs equation is estimated using disaggregated data on transportation costs from the International Transport Database (BTI). Secondly, the effect of transport costs on disaggregated imports is investigated using a direct and an indirect method. The direct method consists of the addition of freight costs and infrastructure variables in the traditional gravity model of trade. The indirect method consists of the separate estimation of an augmented transport-costs equation and a traditional gravity model of trade and combination of the coefficients obtained to calculate the elasticity of trade with respect to transport costs. The results of the indirect method show greater elasticities of trade with respect to transport costs than the direct method (3.4 on average versus 0.85). The primary results highlight the importance of improving transport conditions in order to foster international trade. I. Introduction Recent liberalization processes in Latin America have reduced both tariff and non-tariff barriers considerably. In 1997 Latin America had an average tariff of 8.4 per cent for imports from the EU, considerably lower than the 34-per cent average of 1986.1 This reduc*
1
I would like to thank the participants of the International Economic Workshop, EU and Emerging Markets, held in Vienna, for their helpful comments and suggestions. Financial support from Fundación Caja Castellón-Bancaja and the Spanish Ministry of Education is gratefully acknowledged (P1-1B2005-33 and SEJ 2007-67548). See IMF data.
112
Inmaculada Martínez-Zarzoso
tion in artificial trade barriers implies that transport costs have become an increasingly important determinant of trade. In 1998 the average freight rate for Mercosur imports from the EU was 15 per cent.2 Therefore, any additional effort to integrate a country into the trading system has to consider transport costs. In most cases we have no direct way of observing trade barriers, and therefore have to rely on proxies, indirect measurement, and trade modeling in order to assess the relevance of such barriers. A common practice in the literature related to gravity equations has been to proxy transport costs with the geographical distance between countries. It has been shown that distance seems to be a good proxy for transport costs at an aggregate level. However, it is not clear that this good performance persists at more detailed levels (Hummels, 1999). In this work, we use CIF/FOB ratios obtained from the BTI for the year 1998, including exports from the EU to Argentina, Bolivia, Brazil, Chile, and Uruguay. (Data from the BTI for freight rates are exclusive of loading costs.) This paper is an extension of Martínez-Zarzoso and Suárez Burguet (2005) in which the relationship between trade and transport costs was studied at an aggregated level. We introduce two significant novelties. First, we use disaggregated data on freight rates and estimate the incidence of freight rates for eight sectors (SITC at one-digit level). Rauch (1999) emphasized that transport costs depend upon the type of product traded. He showed that search costs are higher for trade in differentiated products. When aggregated data is used and the composition of trade is not considered, the effect of transport costs (related to distance) in trade patterns can be misleading. To our knowledge, only a few authors have considered disaggregated data, but for different purposes than ours (e.g., Hummels, 1999; Rauch, 1999). Second, we investigate the effect of the transport-costs variable on disaggregated imports by using a direct and indirect method. The direct method consists of the addition of a freight- costs variable in the traditional gravity model of trade. The indirect method consists of the separate estimation of an augmented trade-cost equation and a gravity model of trade (now without the freight-cost variable) and combining the coefficients obtained to calculate the elasticity of trade with respect to transport costs. The augmented trade-cost equation adds infrastructure, port efficiency, and economies of 2
See Database for International Transport (BTI).
On Transport Costs and Sectoral Trade
113
scale as explanatory variables. The estimation of the gravity model with disaggregated bilateral imports will allow us to explore to what extent transport costs have a different effect on imports. Section II provides an overview of the related literature. Section III addresses the difficulties encountered in measuring transportation costs. Finally, Section IV presents the model specifications and section V discusses the results. A concluding section follows. II. The Cost of Moving Goods Across Countries As mentioned in the introduction, distance between countries has been commonly used in the empirical trade literature as a proxy for transport costs. Several authors have recently tried to explain what is behind the so-called “distance effect.” Most of them argue that distance is not only a proxy for transport costs, but also for familiarity, tastes, and cultural similarities. In this line of reasoning, Venables (2001) classifies the costs of distance into four types. The first component is the cost of moving goods internationally (direct shipping costs). The second component is the cost of identifying potential trading partners, sometimes called search costs. The third component includes control and management costs and, finally, the fourth is the cost of time involved in shipping goods. Hence, when distance is added as a regressor in gravity models of trade, it is far from clear how the magnitude of the estimated coefficient should be interpreted. Most research situates the elasticity of aggregated trade flows with respect to distance in the interval (-0.5, -1.5).3 This indicates a major decrease in trade volumes with higher distance. The study of the "new economics of distance" has recently attracted growing interest. Researchers argue that most distance-related costs tend to decrease with the continuous development of new technologies. However, the empirical literature on gravity models shows that the estimated coefficients of distance have been remarkably stable over time. An exception is Coe, Subramanian and Tamirisa (2002). In their non-linear specification of the gravity model, the coefficient estimates of distance decline slightly over time. The cost of moving materials using different modes has changed at different rates across categories without always showing the expected decrease over time. The cost of moving information has
3
Coe et al. (2002, p.5) report a summary of estimates of distance coefficients obtained recently by different authors.
114
Inmaculada Martínez-Zarzoso
indeed steadily declined (92 per cent between 1960-1998)4, but the cost of moving goods has not declined continuously (e.g., the cost of sea transport declined during the 1940's and 1950's but since then, there has been no clear declining trend (Hummels, 2001). Rauch (1999) identified some of the non-conventional costs of trade associated with searching for the case of different types of goods. He claimed that these informational costs fall relatively on differentiated goods. He focused on the effects of distance and common language/colonial ties in trade of three types of goods: those traded on organized exchanges, those possessing reference prices, and all other commodities. His results indicate that proximity and common language/colonial ties are more important in matching international buyers and sellers for differentiated products than for products traded on organized exchanges. Hummels (1999) classified the trade costs implied by trade flows into three different categories: explicitly measured costs given by tariffs and freight rates; costs associated with common proxy variables such as distance, sharing a language, sharing a border, or being an island; and implied but unmeasured trade costs, circumscribed by geographical position, cultural ties, or political stability. According to his results, explicit measured costs are the most important component. He offered alternative explanations for the costs associated with distance and adjacency effects based on direct trade barriers, on endogenous production responses, and on preferences. However, the regression results do not allow to disentangle between various interpretations. Hummels (1999) found that adjacency and distance effects were rarely relevant when interpreted directly as trade barriers. However, when interpreted as price premium, these effects were important in approximately 75 per cent of the products. Moreover, the author gave a tentative interpretation of the residuals of the import-demand equations in terms of the willingness to pay for preferred varieties, assuming that they are unmeasured trade costs. As he pointed out, this interpretation implies that the estimation suffers from omitted variables and also from the biases these variables may induce in the estimated coefficients. Hummels concluded that freight rates at a disaggregated level are higher than aggregated rates and vary considerably among exporters and among products, which suggests that import choices are made in order to minimize transport costs. He also claimed that 4
Ward, Bhattarai and Huang (1999).
On Transport Costs and Sectoral Trade
115
the channels through which trade barriers affect trade volumes remain unclear and are a subject for further research. Finally, he offered a complete characterization of the trade costs implied by trade flows, by establishing a fresh classification of those costs into the three above-mentioned components. In the same vein as Hummels, we focus in the next section on the measurement of transport costs and the analysis of the data used in this paper. III.
Measurement of Transport Costs
One of the main difficulties in analyzing transport costs is obtaining reliable data. In the recent economic literature there have been several attempts to measure transport costs directly or indirectly. Some authors used CIF/FOB5 ratios as a proxy for shipping costs (Baier and Bergstrand, 2001; Limao and Venables, 2001; Radelet and Sachs, 1998). Since most importing countries report trade flows inclusive of freight and insurance (CIF) and exporting countries report trade flows exclusive of freight and insurance (FOB), transport costs can be calculated as the difference of both flows for the same aggregate trade. However, Hummels (2001) showed that importer CIF/FOB ratios constructed from IMF sources are poor proxies for cross-sectional variation in transport costs and such a variable provides no information about the time-series variation. Ogueldo and Macphee (1994) also doubted the usefulness of CIF/FOB ratios from IMF sources as a proxy of transportation costs. Hummels (1999, 2001) used data on transport costs from various primary sources including shipping price indices obtained from shipping trade journals (Appendix 2 in Hummels, 2001), air freight prices gathered from survey data, and freight rates (freight expenditures on imports) collected by customs agencies in the United States, New Zealand, and five Latin-American countries (Mercosur countries plus Chile). In addition to CIF/FOB ratios reported by the IMF, Limao and Venables (2001) used shipping company quotes for the cost of transporting a standard container (40 feet) from Baltimore to 64 destinations. [The authors pointed out that it is not clear how the experience of Baltimore generalized. Martínez-Zarzoso, PérezGarcía and Suárez-Burguet (2003) used data on transportation costs 5
CIF stands for "cost, insurance, and freight"; FOB stands for "free on board."
116
Inmaculada Martínez-Zarzoso
obtained from interviews with logistic operators in Spain. They found import elasticities with respect to transport costs similar in magnitude to those found by Limao and Venables (2001). Micco and Perez (2001) used data from the U.S. Import Waterborne Databank (U.S. Department of Transportation), where transport cost is defined as "the aggregate cost of all freight, insurance, and other charges (excluding U.S. import duties) incurred in bringing the merchandize from the port of exportation to the first port of entry in the U.S." Sánchez, et al. (2003) analyzed data on maritime transport costs obtained from the International Transport Database (BTI). They focused on Latin American trade with NAFTA. The above-mentioned research shows that the common perception that transport costs are unimportant is wrong; they are neither small nor uniform across goods. In the empirical application of this paper, we aim to add further evidence concerning the importance of transport costs. In order to do so, we use CIF/FOB ratios obtained from the BTI.6 (Data from the BTI for freight rates are exclusive of loading costs.) The main difference between these ratios and those reported by the IMF is that the BTI data on imports at CIF prices and imports at FOB prices are obtained from the same reporting country. Since information is collected using identical methodology, the data are more reliable than the IMF rates. A second advantage is that we have disaggregated data at a three-digit level (SITC). Our data are also similar to those used by Hummels (1999), collected by custom agencies in five Latin-American countries (Argentina, Brazil, Chile, Paraguay and Uruguay) in 1992 and 1994. The main differences between both data sets are the period and the set of exporters. We have data for the year 1998 which include imports from the EU to Argentina, Bolivia, Brazil, Chile, and Uruguay (Bolivia is added and Paraguay excluded with respect to Hummels's work). Table A.2 in the appendix shows the average freight rates for each one-digit category. IV.
Model Specification
This section presents the specifications of the transport costs and trade equations and describes the variables used in the empirical application.
6
Data was provided by Jan Hoffmann (ECLA).
On Transport Costs and Sectoral Trade
117
A general formulation of transport costs for commodity k shipped between countries i and j, in a given period of time, can be written as: TCijk = F(Xi, Xj, vi, xijk, Pk, Kijk ) (1) where Xi and Xj are country-specific characteristics, vij is a vector of characteristics relating the journey between i and j, xijk a vector of characteristics depending on the country of origin and destination and the type of product k, Pk is a product-specific effect that captures differences in transport-demand elasticity across goods, and Kijk represents unobservable variables. Among the country characteristics Xi and Xj, we incorporate geographical and infrastructure measures. Typically, dummy variables are used to control for a country that is landlocked or an island. The infrastructure variable used is constructed as an index with larger values of the index indicating a better infrastructure. In the vector of characteristics, vij, distance between trading countries, volume of imports that goes through a particular route, and dummy variables for common language and common border7 are usually considered. Among the characteristics depending also on the type of product, xijk, we focus on the weight value for product k transported from Country i to Country j. Product-specific dummy variables are also modeled to account for Pk. Assuming a multiplicative form, a transport-cost function can be written as:
TCijk
WijkD1 DijD 2 QijD3 InfiD 4 Inf jD5 e
E1 Landi E 2 Land j E3 Langij P k
(2)
where TCijk denotes freight ad-valorem rates, i denotes the importer country, j denotes the exporter country, and k is the three-digit level of the SITC classification. Wijk denotes the weight-to-value ratio (the inverse of the unit value: tons per US$), Dij denotes distance, and Qij is the volume of imports between Countries i and j. This variable accounts for potential economies of scale. Infi and Infj denote infrastructure of Countries i and j. Landi and Landj are dummies that take the value one when the importer or the exporter is a landlocked country, zero, otherwise. Langij takes the value one when Countries i and j speak a common language, zero, otherwise. Kijk denotes the error term that is assumed to be independently dis7
The common border variable is not modelled since in mutual trade, EU-LA countries do not share borders. The same applies for the island dummy.
118
Inmaculada Martínez-Zarzoso
tributed. The variables, Infi and Infj, are first constructed as an index8 (taking information on roads, paved roads, railroads, and telephones), by differentiating between importer and exporter country infrastructure as explanatory variables of transport costs. Our index is similar to that of Limao and Venables (2001). A rise in our index indicates better infrastructure and is expected to be associated with a decrease in the costs of transport. We also used data on port efficiency from the Global Competitiveness Report as an alternative for the infrastructure variables. These data consist of an index taking the values 1 to 7, with 7 indicating the best score. We estimate a linear version of Equation 2. Taking natural logarithms, the general specification is given by ln TCijk
D 1Wijk D 2 ln Dij D 3 ln Qij D 4 ln Inf i D 5 ln Inf j
E1 ln Land i E 2 ln Land j E 3 Lang ij P k K ijk
(3)
where ijk denotes the error term which is assumed to be i.i.d.. To assess the relative importance of transport costs on trade, a gravity equation augmented with infrastructure and transport costs variables is estimated. According to the simplest gravity model of trade, the volume of aggregated imports (exports) between pairs of countries, Mij, is a function of their incomes (GDPs), their geographical distance, and a set of dummies: M ij J 0Yi J 1 Y jJ 2 D ijJ 3 AijJ 4 u ij (4) where Yi (Yj) indicates GDPs of the exporter (importer), Dij measures the distance between the two countries’ capitals (or economic centers), and Aij represents any other factors aiding or preventing trade between pairs of countries. uij is the error term. Trade is expected to be positively related to economic mass and negatively related to distance. The majority of estimates of the gravity equation are based on a log-linear transformation of different versions of Equation 4. Considering now sectoral trade, the linear version is given by,
8
Infrastructure in each country is measured by an index constructed by taking the mean of four variables; km of road, km of paved road, km of rail (each one divided by the surface area of the country) and main telephone lines per person. Since these measures are highly correlated among themselves, it is not possible to identify each of their influences on transport costs separately.
On Transport Costs and Sectoral Trade
ln M ijk
119
J 0 J 1 ln Yi J 2 ln Y j J 3 ln Dij ¦ G h Pijh H ijk , (5) h
where ln denotes variables in natural logs. ¦ G h Pijh is a sum of h
preferential-trade dummy variables. Pijh takes the value one when a certain condition is satisfied (e.g., belonging to a trade bloc), zero, otherwise. Usually models include dummy variables for trading partners sharing a common language and common border, as well as dummy variables for trading blocs evaluating the effects of preferential trading agreements. The coefficients of all these trade variables (Gh) are expected to be positive. Population or income-percapita variables are usually added to the list of explanatory variables to model absorption and economies-of-scale effects. We added infrastructure variables instead. Since per capita incomes are highly correlated with infrastructure endowments, the two variables could not be added simultaneously to the estimated equation. Distance is also included in most empirical gravity models. However, the theoretical models relate trade to transport costs, for which distance is a proxy. The relationship between trade cost and distance, already specified in Equation 3, above, indicates that the elasticity of trade costs with respect to distance is positive, so D2>0. Considering a theoretical model that relates bilateral trade at a disaggregated level with income and transport costs and then substitutes Equation 3 for transport costs, one obtains an augmented gravity equation relating trade to distance and other variables: ln M ijk
J 0 J 1 ln Yi J 2 ln Y j J 4 ln TCijk H ijk
(6)
where ln TC ijk
D 1Wijk D 2 ln Dij D 3 ln Qij D 4 ln Inf i D 5 ln Inf j
E1 ln Land i E 2 ln Land j E 3 Lang ij P k K ijk
(7)
In the standard model (Equation 5 without dummies), J3 is a “gross” distance elasticity of trade, while in the models augmented with infrastructure or/and dummies and once controlled for the composition of trade (sectoral dummies), the distance elasticity is a “residual.” The model is estimated for bilateral exports from twelve EU countries to five Latin-American importers with data for 1998 disaggregated at a three-digits level (SITC). We performed OLS estimation on the double-log specification as given by Equations 3, 5, and 6.
120
Inmaculada Martínez-Zarzoso
V. Results
Table 1 presents the results of estimating Equation 3 at one-digit SITC level. We found that there is heterogeneity among the estimated coefficients when aggregated over all products, especially concerning the effect of distance. In fact, we observe important differences across sectors. The weight-value ratio, landlocked dummy, and importer infrastructure variables are significant at the one-per cent level for all categories. However, the distance variable is only significant for sectors 7 and 8 when infrastructure variables are added. The magnitude of the product-weight elasticity with respect to transport costs varies within the range (0.12-0.37). Sectors 1, 7, and 8 show the highest elasticity. Those are sectors with higher than average unit values and therefore, air transport may be a good substitute for sea transport. Additionally, the significance and magnitude of the distance elasticity is higher for manufactured products (with the exception of sector 3). Products in these sectors are the most heavily exported from the EU to Latin America. When the landlocked dummy is added as a explanatory variable, it shows a significant and positive coefficient within the range (0.30-0.95) which confirms the higher transport costs incurred for landlocked countries (in our case, Bolivia). Transport costs for landlocked countries is between 35 per cent [exp(0.30)-1] and 158 per cent [exp(0.95)-1] higher than for non-landlocked countries. Table 2 presents the results of estimating Equation 6 at the onedigit SITC level. Two versions are estimated, one with the landlocked dummy and the second with infrastructure variables (importer infrastructure was highly correlated with the landlocked dummy, hampering the estimation results of the equation with both variables added simultaneously). In both cases we can see that the fit of the regressions is better for manufactured products (5-8) and that the significance and sign of the distance and infrastructure variables vary considerably across sectors. In most cases the infrastructure variables exert a positive effect on trade. The average effect indicates that a 10-per cent improvement in the infrastructure of the importer will increase trade by about 25 per cent. Income elasticities are lower for the exporters (European countries) than for the importers (Latin-American countries) and more homogeneous products also present lower income elasticities. Income per capita variables were not included because they where highly correlated with the infrastructure indices and also with the
On Transport Costs and Sectoral Trade
121
landlocked dummy. The distance coefficient varies greatly across sectors in magnitude (from 0.07 to 5.80); it shows the expected negative sign in seven out of nine cases, and only for six sectors is the coefficient statistically significant. Table 3 shows direct estimates of export elasticities with respect to transport costs. Equation 5 is estimated adding distance, as well, and the landlocked and language dummies to the specification and considering transport costs as exogenously determined. The results show that the transport-cost variable is always statistically significant and presents the expected negative sign in all but one of the estimations (Sector 3). The range of the direct elasticities is (-0.30, 1.57). In contrast, distance is in five cases insignificant and presents an unexpected positive sign in five cases. We found that we cannot accept the equality of the distance coefficient at the one-digit level and this perhaps may be the reason for these results. However, we can accept the equality of the transport costs coefficients when allowing for heterogeneity at the two-digit level. We also estimated a system of equations simultaneously (Equations 6 and 7). The main problem was that most of the determinants of transport costs were also related to trade and the possible instruments for transport costs could also enter directly in the trade equation. As an alternative, to overcome the problem of endogeneity of the transport-costs variable in the gravity equation, we estimated transport costs and trade equations separately and then used the coefficients of some of the explanatory variables that enter both equations to indirectly calculate the elasticity of trade with respect to transport costs. In Table 4 the indirect method is used to calculate the elasticity of trade with respect to transport costs. The variables used are the distance and the landlocked dummy. The elasticity is calculated dividing the coefficient in the trade equation by the coefficient in the transport costs equation for a given variable. Considering only the results based on statistically significant coefficients as valid, the elasticity of trade with respect to transport costs varies within the range (-1.04, -8.57) when the calculus is based on the landlocked dummy and within the range (-2.20, -5.63) when based on the distance variable.
0.38 (0.80) -0.05 (-0.12) 0.281 (1.00) -0.32 (-1.09) 1.208 (2.46) *** 0.838 (1.59) 0.19 (0.36) -0.29 (-0.25)
(5.17)*** (5.52)***
0.36 0.37
0.258 (8.89)*** 0.263 (9.13)***
0.191 (4.72)*** 0.183 (4.49)***
0.127 (2.10) *** 0.127 (3.28) ***
0.239 (14.60) *** 0.244 (14.94) ***
0.204 (14.28) *** 0.207 (14.18) ***
0.276 (19.75) *** 0.283 (20.03) ***
0.343 (19.55) *** 0.356 (19.81) ***
-3.32 (-0.74) -0.59 (-0.13)
-3.45 (-1.32) 0.59 (0.22)
-11.13 (-2.45) -8.59 (-1.82)
-2.93 (-0.59) 0.12 (0.02)
-6.64 (-4.81)*** -3.13 (-2.29)**
-5.02 (-4.96)*** -2.09 (-2.00)**
-4.63 (-4.85)*** -2.86 (-2.90)***
-4.28 (-3.04)*** -2.97 (-2.09)***
1
2
3
4
5
6
7
8
-
0.402 (7.07) ***
-
0.405 (9.36) ***
-
0.55 (10.47) ***
-
0.51 (8.19) ***
-
0.46 (2.97) ***
-
0.526 (2.50) ***
-
0.95 (7.14) ***
-
0.30 (1.85) ***
-
Landlocked country dummy 0.478 (4.58) ***
-0.18 (-4.61)*
-
-0.23 (-7.85) ***
-
-0.27 (-7.61) ***
-
-0.36 (-8.28) ***
-
-0.42 (-2.92) ***
-
-0.27 (-1.80) ***
-
-0.50 (-5.50) ***
-
-0.45 (-2.88) *** **
-
-0.36 (-5.42) ***
-
Own infrastr.
0.03 (0.88)
-
-0.01 (-0.53)
-
-0.09 (-2.98) ***
-
-0.14 (-3.63) ***
-
-
-
-
-
-
Partner infrastr.
0.449 0.925
0.321 0.309
0.273 0.250
0.394 0.397
0.194 0.221
0.438 0.420
0.338 0.314
0.234 0.271
Adj. R2 0.313 0.331
1108
2002
1690
1164
95
96
390
98
670
Nobs
Note: Triple, double, and single asterisks denote significance level of 1, 5, and 10 per cent, respectively. T-values based on heteroskedasticity-consistent standard errors are in brackets.
0.538 (3.53) *** 0.345 (2.21) ***
0.473 (4.62) *** 0.20 (1.89)**
0.483 (4.39) *** 0.066 (0.57)
0.61 (1.08) *** 0.11 (0.72)
0.50 (2.68)** 0.128 (0.65)
-3.29 (-1.93) -0.98 (-0.57)
Distance to exporter
0
Weight/value (kg/$) 0.232 (5.54)*** 0.234 (5.53)***
Constant
Sectors
Table 1: Estimation Results: Sectoral Structure of Freight Rates for Five LA Importers in 1998
1.37 (24.32) *** 0.07 (0.16) 1.46 (24.62) *** -0.21 (-0.50) 1.22 (16.40) *** -3.52 (-5.60) *** 1.21 (15.04) *** -1.91 (-3.33) ***
0.05 (0.35) 0.29 (2.42) ***
1.05 (2.539) *** 1.20 (3.84) ***
0.33 (1.00) 0.36 (1.51)*
0.64 (8.58) *** 0.82 (14.61) ***
0.48 (7.97) *** 0.84 (14.96) ***
0.66 (12.32) *** 0.71 (16.92) ***
0.30 (4.70) *** 0.54 (10.45) ***
28.45 (1.68)* 4.99 (0.35)
-81.36 (-1.82)** -101.5 (-2.73) ***
-5.31 (-0.11) -2.66 (-0.07)
-34.77 (-3.96) *** -54.00 (-7.56) ***
-15.88 (-2.46) *** -33-64 (-5.91) *****
-51.72 (-8.86) *** -52.61(-10.58) ***
-5.14 (-0.96) -25.13 (-3.88) ***
3
4
5
6
7
8
-1.51 (-5.38) ***
-
-0.42 (-1.69)*
-
-1.68 (-6.09) ***
-
-1.32 (-3.82)*
-
0.23 (0.16)
-
-2.13 (-2.89) *** -1.81 (-0.99)
-
-2.57 (-2.26) ***
-
-1.17 (-2.57)
***
0.22 (1.42)*
-
0.06 (0.49)
-
0.31 (2.07)**
-
0.42 (2.36) ***
-
-0.31 (-0.41)
-
0.91 (0.93)
-
0.48 (1.23)
-
1.18 (1.94)**
-
0.26 (1.08)
-
Landlocked dummy Own infrastr.
-
-0.64 (-5.68) ***
-
-0.31 (-2.23)**
-
0.64 (4.01) ***
-
0.60 (0.94)
-
-0.45 (-0.50)
-
0.003 (0.01)
-
0.86 (1.58)*
-
0.467 (1.93) ***
-
Partner infrastr.
0.381
0.390 0.399
0.351 0.341
0.354 0.357
0.016 0.015
0.368 0.362
0.109 0.091
0.239 0.240
0.176 0.172
Adj. R2
0.18 0.361 (1.15) Note: Triple, double, and single asterisks denote significance level of 1, 5 and 10 per cent, respectively. T-values based on heteroskedasticity-consistent standard errors are in brackets. Two estimations for each sector are presented. In the first row, the landlocked dummy is included as a regressor and in the second, this dummy is dropped and infrastructure data of the exporter and importer countries are added.
1.07 (15.80) *** -2.72 (-5.11) *** 1.125 (15.15) *** -1.48 (-2.97) ***
1.30 (17.08) *** -1.34 (-1.83)** 1.24 (16.09) *** 0.57 (0.35)
1.69 (0.46) 1.47 (0.49)
(-1.48) (-1.72)*
-0.58 -0.76
-3.89 (-2.98) *** -1.99 (-1.70)*
(4.93) *** -0.14 (-0.04) (4.98) *** 1.78 (0.57)
(2.04)** (2.13)**
2.21 2.26
0.36 0.40
(2.69) *** -5.80 (-2.47)* (1.93)** -2.96 (-1.48)*
2
0.67 0.53
-0.03 (-0.17) 0.20 (1.09)
-1.60 (-1.86) -0.38 (-0.49)
41.55 (1.67) 17.31 (0.77)
(1.68) (1.26)
**
Distance to exporter
1
0.19 0.15
*
***
0.349 (3.66) 0.512 (6.80) ***
Exporter GDP
Importer GDP
2.26 (0.22) -11.30 (-1.27)
Constant
0
Sect.
Table 2: Sectoral Estimates of Import Demand
Importer GDP
Exporter GDP
Distance to exporter
Landlocked dummy
Transport costs
Lang.
Adj. R2
0
Yes
0.30*** 0.60*** 0.95 -0.58 -0.52*** 0.92*** 0.26 (7.54) (3.29) (1.16) (-1.43) (-4.79) (3.59) 1 Yes 0.27 0.59 -3.10* -0.86 -1.57*** 0.34 0.36 (1.24) (2.78) (-1.59) (-0.76) (-5.65) (0.53) 2 Yes 0.30*** 0.28*** -1.40*** -0.30*** 0.59** 0.25 -1.58* (2.60) (-1.60) (-2.50) (-2.59) (1.92) (2.23) 3 Yes 0.90*** 0.59 0.40 1.44*** 0.60 -1.38 0.80** (2.09) (0.97) (3.54) (5.00) (0.26) (-1.17) 0.09 3.70* 0.96 0.14 4 Yes 0.77*** 1.64 -1.17*** (2.99) (0.92) (-3.08) (1.22) (1.41) (1.52) 1.07*** 5 Yes 0.80*** 0.76 -0.62*** -0.72*** 0.81*** 0.54 (14.68) (20.06) (1.41) (-2.52) (-7.96) (5.04) 6 Yes 0.63*** 0.97*** -0.12 -1.22*** -0.69*** 1.15*** 0.45 (13.64) (19.01) (0.27) (-5.46) (-8.66) (7.46) 7 Yes 0.85*** 2.14*** 0.03 -0.94*** 1.02*** 0.51 1.18*** (26.07) (4.91) (0.15) (-12.31) (7.00) (19.32) 8 Yes 0.51*** 1.14*** -0.88*** 1.35 0.53 -0.35 -1.34*** (9.14) (-5.66) (-10.04) (7.99) (19.03) (-0.63) Note: Triple, double and single asterisks denote significance levels of 1, 5 and 10 per cent, respectively. T-values based on heteroskedasticity-consistent standard errors are in brackets. Only the specification with the landlocked variable is estimated. Infrastructure variables are not added because they are correlated with the transport costs variable.
Sect. Sectoral dummies
Table 3: Direct Estimates of Export Elasticity with Respect to Trade Costs
Elasticities Transport cost equation Trade equation Export elasticities w.r.t. transport costs Sector Distance Distance 0 0.5 -1.16 -2.32 1 0.38 ns -2.96 -7.79 2 0.28 ns -1.99 -7.11 3 1.2 ns -0.14 ns -0.12 4 0.19 ns 1.69 ns 8.89 5 0.61 -1.34 -2.20 6 0.483 -2.72 -5.63 7 0.473 -0.21 ns -0.44 8 0.538 -1.91 -3.55 landlocked landlocked Sector 0 0.478 -1.17 -2.45 1 0.3 -2.57 -8.57 2 0.95 -2.13 -2.24 3 0.526 -1.81 ns -3.44 4 0.46 0.23 ns 0.50 5 0.51 -1.32 -2.59 6 0.55 -1.68 -3.05 7 0.405 -0.42 -1.04 8 0.402 -1.51 -3.76 Note: The point estimates for distance and landlocked variables in the transport cost equation, are from Table 1. The point estimates for distance and landlocked variables in the export equation are from Table 2. ns denotes not significant at conventional levels .
Table 4: Estimates of Export Elasticity with Respect to Transport Costs-Indirect Method
126
Inmaculada Martínez-Zarzoso
When these findings are compared with other studies,9 the elasticities found in this paper are slightly higher, which is to be expected since transoceanic shipments involve extra costs. The elasticities of trade with respect to transport costs obtained using the indirect method show a higher magnitude than those obtained using the direct method. This outcome could indicate that the transport costs’ effect on trade is underestimated when we fail to account for possible reductions in transport costs stemming from economies of scale in transport. Sectors with higher trade volumes could profit from lower transport costs. An average reduction in freight rates of one per cent will give rise to an increase in the trade of food and live animals (Sector 0) of about two per cent, whereas it produces an increase in trade of beverages and tobacco (Sector 1) of around seven per cent (Table 4). The machinery and transport equipment categories show the lowest effect (1 per cent). When we compare these elasticities with the elasticity of trade with respect to distance, we observe that the distance effect decreases for most sectors when freight rates are added as an explanatory variable in the gravity equation. However, although distance can partially account for the effect of trade costs, it is still significant and shows a negative effect for beverages and tobacco and crude materials. For Sectors 4 and 7 (oils and fats and transport equipment), we obtain an unexpected positive coefficient, that is not significant, for the distance variable when freight costs are explicitly added as a regressor. It could be that for those sectors and for some reasons (e.g., comparative advantage) the importers trade more with countries that are more distant. VI.
Conclusions
The objective of this paper was to investigate the determinants of transport costs and the nature of the relationship between trade and transport costs for Latin-American sectoral imports from the European Union. We first estimated a transport-costs equation using data on transportation costs. We then analyzed the relationship between transport costs and trade using a direct and an indirect method. The results from our first estimation show that higher distance and poor partner infrastructure notably increase transport costs. In9
Martínez-Zarzoso et al. (2007) showed elasticities in the range (-1.20,3.93) for Spanish exports to Europe. See Table A.1 in the appendix.
On Transport Costs and Sectoral Trade
127
clusion of infrastructure measures improves the fit of the regression, corroborating the importance of infrastructure in determining transport costs. The distance coefficient loses significance and decreases in magnitude when we add infrastructure variables. An improvement of 10 per cent in the infrastructure of the importer country leads to a reduction in transport costs of 35 per cent on average for Sectors 0 through 5 and of 22 per cent on average for Sectors 6 through 8. The gravity model estimates show that the elasticity of the importer and exporter income variables are lower than unity for more homogeneous products and have, as expected, a positive influence on bilateral trade flows. Importer income elasticity is lower than exporter income elasticity for all sectors, reflecting the exportoriented structure of most European Union countries. Being landlocked significantly deters trade, whereas sharing a language fosters trade. Infrastructure improvements have a positive effect on sectoral trade, but the effect is only statistically significant for half of the sectors. Whereas we find clear evidence showing a direct effect of infrastructure improvements on the reduction of transport costs, the effect of infrastructure on trade is less pronounced. We find the expected negative effect of distance on trade for six sectors out of nine. When the gravity model is augmented with the transport-costs variable, the results show that in only one sector (mineral fuels and lubricants) out of nine there is an unexpected positive coefficient. The average elasticity of trade with respect to transport costs is 0.85 following this direct method. Distance loses significance and the coefficients decrease in magnitude. The results from the indirect method based on separated estimation of the transport costs and the trade equations and using the results to compute the elasticity, show higher elasticities of trade with respect to transport costs than do the results of the direct method (3.4 on average when using distance as the common regressor in both equations). This fact could be indicative of the importance of improving transport conditions in order to foster international trade. Finally, our results are more robust for Sectors 5 through 8 than for the others, given that we have a higher number of observations.
128
Inmaculada Martínez-Zarzoso
References
Coe, David T., Subramanian, Arvind, Tamirisa, Natalia T. (2002), The missing globalization puzzle, International Monetary Fund Working Paper WP/02/171, Washington. Deardorff, Alan (2001), Local comparative advantage, trade costs and the pattern of trade, University of Michigan Discussion Paper 500. Baier, Scott L., Bergstrand, Jeffrey H. (2001), The growth of world trade: Tariffs, transport costs, and income similarity, Journal of International Economics 53, 1-27. Fink, Carsten, Mattoo, Aaditya, Neagu, I. Cristina (2000), Trade in international maritime service: How much does policy matter?, World Bank Economic Review 16 (1), 81-108. Hoffmann, Jan, Kumar, Shashi (2002), Globalization—The maritime nexus, in: Grammenos, Costas. T. (ed.), Handbook of Maritime Economics and Business, LLP, London, Chapter 3, 3562. Hummels, David (1999), Towards a geography of trade costs, University of Chicago, Mimeographed document. Hummels, David (2001), Have international transportation costs declined?, Journal of International Economics 54 (1), 75-96. Limao, Nuno, Venables, Anthony J. (2001), Infrastructure, geographical disadvantage and transport costs, World Bank Economic Review 15 (3), 451-479. Martínez-Zarzoso, Inmaculada, García-Menendez, Leandro, Suárez-Burguet, Celestino (2003), The impact of transport costs? on international trade: The case of Spanish ceramic exports, Maritime Economics and Logistic 5 (2), 179-198. Martínez-Zarzoso, Inmaculada, Suárez-Burguet, Celestino (2005), Transport costs and trade: empirical evidence for LatinAmerican imports from the European Union, Journal of International Trade and Economic Development 14 (3), 227-245.
On Transport Costs and Sectoral Trade
129
Martínez-Zarzoso, Inmaculada, Pérez-García, Eva, Suárez-Burguet, Celestino (2008), Do transport costs have a differential effect on trade at the sectoral level?, Applied Economics, forthcoming. Micco, Alejandro, Pérez, Natalia (2001), Maritime transport costs and port efficiency, Inter-American Development Bank, Research Working Paper 441. Oguledo, V.I. and Macphee, C.R. (1994), Gravity Models: A Reformulation and an Application to Discriminatory Trade Arrangements, Applied Economics 26, 107-120. Radelet, Steve, Sachs, Jeffrey (1998), Shipping costs, manufactured exports and economic growth, Harvard University, Harvard Institute for International Development, Mimeographed document. Rauch, Jaames E. (1999), Networks versus markets in international trade, Journal of International Economics 48, 7-35. Sánchez, Ricardo J., Hoffmann, Jan, Micco, Alejandro, Pizzolitto Georgina, Sgut, Martin, Wilmsmeier, Gordon (2003), Port efficiency and international trade: Port Efficiency as a Determinant of Maritime Transport Costs, Maritime Economics and Logistics 5 (2), 199-218. Venables, Anthony J. (2001), Geography and international inequalities: The impact of new technologies, Journal of Industry, Competition and Trade 1 (2), 135-159. Ward, w. A., Bhattarai, M. and Huang, P. ((1999), The new economics of distance: Long Term Trends in Indexes of Spatial Friction, Department of Agricultural and Applied Economics Working Paper 020299, Clemson University.
E2
Trade equation
Dist. Dist. Agro-industry 0.61 -0.94 Ceramics 0.69 -0.84 Household app. 0.46 -1.31 Vehicle parts 0.31 -1.22 Note: Indirect method. Table 3 in Martínez-Zarzoso, et al. (2007).
D1
Transport cost equation
E 2 D1
-1.52 -1.20 -2.82 -3.93
W
Export elasticities
Table A.1: Estimates of Export Elasticity with Respect to Transport Costs for Spanish exports to Europe
Appendix
SITC
TC-AR
TC-BR
TC-CH
0 Food and live animals 0 0.19 0.18 0.21 1 Beverages & tobacco 1 0.12 0.12 0.15 2 Crude materials 2 0.13 0.21 0.28 3 Mineral fuels, lubricants 3 0.31 0.18 0.27 4 Animal & vegetable oils, fats 4 0.15 0.13 0.12 5 Chemicals & related Products 5 0.11 0.09 0.12 6 Manufactured goods 6 0.1 0.11 0.13 7 Machinery & transport equipment 7 0.09 0.07 0.07 8 Misc. manufactures 8 0.1 0.09 0.11 Source: International Transport Database (ECLAC). TC denotes transport costs.
SITC-2D
Table A.2: List of Sectors and Ad-Valorem Sectoral Freight Rates
0.14 0.07 0.12 0.37 0.08 0.09 0.1 0.08 0.1
TC-UR 0.36 0.15 0.28 0.17 0.16 0.23 0.41 0.15 0.27
TC-Bo 0.27 0.12 0.18 0.28 0.12 0.1 0.11 0.08 0.1
Aver-4
0.22 0.12 0.2 0.26 0.13 0.13 0.17 0.09 0.13
Aver-5
Rodrigo M. Zeidan
Evolving Patterns of Trade between Brazil and the European Union 1994-2005 Abstract The evolution of trade patterns is of major importance for trade negotiations, since changes in trade patterns have important implications for the formulation of trade policies. The goal of this paper is to provide a dynamic test to evaluate the evolution of trade patterns between Brazil and the European Union following Proudman and Redding (2000) in calculating transition probabilities to measure mobility across export sectors. In addition, a structural change test in the sense of Mancini-Griffoli and Pauwels (2006) is used to verify if the patterns of trade are constant in the 1994-2005 period. The results of this analysis reveal a typical North-South trade pattern between the EU and Brazil. Trade specialization is primarily based on factor endowment although some evidence of increasing intra-industry trade and trade diversification appears. I. Introduction Models of North-South trade are only one of the issues of the trade literature. However, due to the growing share of developing countries in world trade, this relatively neglected area becomes more and more important.1 A main interest of North-South trade research is evolvement of trade patterns.2 It is a thriving area of research, since many of the results have direct impact on policy formulation. Although conceptualized for the specific situation of developing countries, models of North-South trade follow regular trade models. Since developing countries are experiencing a late industrialization 1 2
For a survey of North-South trade models see Chui et al. (2002). See for example Bastos and Cabral (2007), Kaplinsky et al. (2006), Zaghini (2005), Uchida and Cook (2005), Stehrer and Wörz (2003), Curtis and Shenjie (2003), Proudman and Redding (2000), Brasili et al. (2000), Labson (1997), Zhang (1994), James (1994).
134
Evolving Patterns of Trade between Brazil and the EU
processes, North-South models usually analyze issues like technological diffusion, productivity links, oligopolistic dynamics etc. Textbook trade models predict that evolving patterns of trade include not only growing specialization depending on factor endowments (inter-industry trade), but also increasing intra-industry trade and trade diversity. Moreover, since trade expansion is linked to the industrialization process, linkages between technology and trade dynamics are fundamental to explain this evolution. The intention of this is to build upon current research in formalizing a test for evolving patterns of trade between Brazil and the European Union. Although econometrically simple, the test can be viewed as a way to test if the evolution of trade between two regions follows a locked-in factor endowment-based trade pattern or if technology and intra-industry play a larger role. There are three main considerations motivating this investigation: First, Brazilian exports show an impressive growth, with an annual growth rate of 15 per cent in 2001-2006. This is the result of a massive currency depreciation in 2002 and a production towards a more export-led growth than experienced by the country previously. Second, considering trade negotiations between Brazil and the EU, formal tests of trade patterns are an important tool to formulate trade policies. Finally, the present trade literature does not yet provide an analysis of the stylized facts that characterize the bilateral trade between the EU and Brazil and its dynamics considering technological content, comparative advantages on a HS06 product level, evolving patterns of trade and so on. Trade relations between the EU and Brazil are an interesting case for analyzing the trade structure and its dynamics. Focusing on bilateral trade also has the advantage of using the stylized facts as an important tool for analysis of the results. Analyzing trade patterns of either Brazil or European Union with the rest of the world would make the results less clear since many different effects would impact the trade flows, such as cross-country trade barriers, changes in tariff structure, FDI, etc. The paper is organized as follows. The second section presents a review on the relevant literature and the models to be tested and analyzed. The third section presents stylized facts and analysis of the trade flows between Brazil and the EU, as well as data construction procedures and the mains results. The final section brings some final comments.
Rodrigo M. Zeidan
135
II. Theoretical arguments and modelling of the evolution of trade patterns That the world trade grows at a briskly pace is hardly news, but is still surprising, since economic theory cannot explain why world trade is growing so fast. A good example can be found in Bergoeing and Kehoe (2001). The authors develop a model that allows changes in relative prices to have large effects, assumes differences in income distribution among industrialized countries, and postulates non-homothetic preferences to affect trade directions and volumes. Analyzing the 1950-1990 period, Bergoeing and Kehoe’s model backwardly predicts positive and large world trade growth, but does not predict the extreme growth of the ratio of trade to world income, from 7.9 per cent in 1950 to 15.4 per cent in 1990. From 1990 on this pattern does not change, with trade in 2006 accounting for almost 25 per cent of global GDP (World Economic Outlook, 2007), a fact that no trade model can explain. Partial explanations include the decrease in average tariff, intraindustry trade, high elasticity of substitution and financial links, among others. The most common explanation is tariff reduction, but tariffs are decreasing at a slower pace than in previous decades, and world trade is still consistently growing more than world income. Yi (2003) develops a model with vertical specialization to try to explain the growth conundrum. The model generates a non-linear response - which may be the best way to explain world trade growth - of trade to tariff reductions, but the model still explains only slightly more than 50 per cent of the world trade growth. Alternative developments look into reduction of non-tariff barriers as a source for the extreme trade growth (Hummels, 1999; Wacziarg and Welch, 2003). Irwin (2002) observes that there is increasing sensitivity of trade to income, and even though trade is growing consistently, its growth pattern is becoming increasingly linked to income growth, with trade following cycles much like income cycles. Also, until the early 1990’s trade growth was skewed towards developed countries, with trade growing more through intra-industry trade between developed countries than through North-South inter-industry trade. This scenario began to change with emerging markets like China, India and Brazil presenting double-digit annual trade growth. In the period 1989-1998, trade growth in developed countries averaged 6.6 per cent, while 7.4 per cent in emerging markets. This differ-
136
Evolving Patterns of Trade between Brazil and the EU
ence is growing, with a 5.6 per cent average annual trade flows growth in developed markets compared to 9.5 per cent in emerging markets in the 1999-2007 period (World Economic Outlook, 2007). As observed by Bergoeing and Kehoe (2001), the new trade theory was developed to account for three major stylized facts: the ratio of trade to income has increased; trade has become more concentrated among industrialized countries; and trade among industrialized countries has been largely intra-industry trade. While the first stylized fact is still true, the others are relatively less important since most of the world trade growth is being carried out by interindustry trade too, and from, emerging markets. Moreover, intra-industry trade is also increasing between advanced and emerging markets. In Brazil, for instance, a sizable part of the trade with the United States is manufactured goods - 20 per cent of the trade flow is concentrated in the HS02 chapter 84 – Machinery and mechanical appliances. Another key interest of the recent trade literature is the link between technology transfer and trade.3 Technology transfer through trade is usually considered to take place via the technological component of imports. In this way, developing countries would adopt technological advance by importing more technology-intensive products and learning by imitation. As Connolly (1997) observes, the process of imitation can lead to dynamic gains of trade to the South since “free trade generally results in a positive feedback effect between Southern imitation and Northern innovation yielding a higher common steady-state growth rate”. Keller (2002) corroborates that those gains are relevant, and develops a framework where “technology in the form of product designs is transmitted to other industries, both domestically as well as internationally, through trade in differentiated intermediate goods”. Lumenga-Neso et al. (2001) makes the interesting case that ‘indirect’ trade-related technology transfer also matters. This ‘indirect’ transfer is defined as one associated with available rather than with produced levels of R&D. The authors conclude that “these ‘indirect’ trade-related spillovers are at least as important as the ‘direct’ ones, and strengthen the view that trade does matter for the international transmission of R&D”.
3
See for example Cameron et al. (2005), Connolly (1997), Hoppe (2005), Lumenga-Neso et al. (2001), Keller (2002).
Rodrigo M. Zeidan
137
Recent developments in econometrics allow for many diverse analyses of trade patterns, with some based on the aforementioned link between factor productivity and trade pattern. Curtis and Shenjie (2003) focus on the impact of trade costs of trade patterns. Sachs et al. (2002, p. 22) develop a model to “provide a unified framework for the analysis of trade patterns and economic development of monopolistic competition.” Trade pattern studies are so diverse that even the role of inequality in trade patterns has been analyzed (Requena-Silvente and Walker, 2007). The present work follows the line of James (1994), Proudman and Redding (2000), Brasili et al. (2000), Bastos and Cabral (2007) and many others, which focus on the dynamics and the evolution of trade patterns. The range of these works goes from mainly descriptive analysis (James, 1994) to a throughout econometric estimation (Bastos and Cabral, 2007), with many intermediate steps. Usually no structural models are used, since it is difficult to derive testable equations that can be analyzed with trade data, especially if a researcher considers changes in factor productivity. The main drive is that factor productivity changes are revealed in the evolution and dynamics of trade patterns, hence, in trade data. We wish to develop a comprehensive analysis of the evolution of patterns of trade between the European Union (EU) and Brazil in the 1994-2005 period and to verify if this evolution is mainly due to inter-industry or intra-industry trade. This comprehensive approach consists of three main lines of research: a descriptive analysis using trade data and indexes; a statistical analysis following Proudman and Redding (2000) and Brasili et al. (2000); and an econometric dynamical test that builds upon Bastos and Cabral (2007), ManciniGriffoli and Pauwels (2006) and other works. The descriptive analysis follows in the next section. The statistical analysis is based on a model of Proudman and Redding (2000), which considers the evolution of patterns of trade based on changes in productivity due to technological changes, with complementary analysis done by Brasili et al. (2000). Proudman and Redding (2000) present a way to analyze the extent to which trade patterns persist over time (what the authors call intra-distribution dynamics). The argument is that the trade literature neither presents a clear argument for mobility nor for the persistence of Revealed Comparative Advantages (RCA) over time. Evolution of trade can be experienced by constant RCA of trading partners or shifting exports and imports across export sections.
138
Evolving Patterns of Trade between Brazil and the EU
Trade models predict that the two happen over time, and the authors try to investigate the degree of persistence versus mobility in patterns of international specialization. The rationale is pretty clear and is summarized by the authors: “much of the existing empirical trade literature is concerned with patterns of international trade at a point in time. This (...) stands in marked contrast with the theoretical literature on growth and trade, which emphasizes that comparative advantage is dynamic and evolves endogenously over time” (Proudman and Redding, 2000, p. 373). The model employed has the purpose of deriving changes in the dynamics of patterns of trade. The static equilibrium is based on a simple Ricardian model with continuous goods. The equilibrium condition is such that a country produces goods following the condition that the country’s relative wage is higher than the country’s productivity relative to the foreign country´s productivity. Two more conditions have to be met: a country’s income equals world expenditure on the country’s goods, and instantaneous utility is a symmetric Cobb-Douglas function. The main result of the static model (specification and notation can be found in Proudman and Redding, 2000) is that the evolution of patterns of international trade over time is determined by rates of technological progress in each section of the two economies (Proudman and Redding, 2000, p. 376). To develop a dynamic framework the authors consider three sources of technological change: learning-by doing (J); knowledge spillovers (\ ); and a constant function based on unobserved variables (O). With Aij as productivity in section j of economy i, Lj as the amount of labor (the only production factor), and Axj as productivity in section j in whichever of the two economies is the technological leader, the rate of productivity growth is a linear additive function of these three sources given by:
§ Aij (t ) · ¸ ln¨ ¨ Aij (t 1) ¸ © ¹
§ Axj (t 1) · ¸, ¨ Aij (t 1) ¸ © ¹
J ij \ j ln(1 Lij (t 1)) O j ln¨ J ij ,\ ij , Oij t 0 i, j
(1)
The evolution of the productivity of each section of each economy relative to the world technological leader is then given by:
Rodrigo M. Zeidan
§ Aij (t ) · ' ln ¨ ¸ ¨ Axj (t ) ¸ © ¹
139
§ 1 Lij (t 1) · (J ij J xj ) \ j ln ¨ ¸ ¨ 1 Lxj (t 1) ¸ © ¹ (2) § Aij (t 1) · O j ln ¨ ¸ ¨ Axj (t 1) ¸ © ¹ The result of the model is that the patterns of trade are initially determined by factor productivity but its evolution over time depends on the dynamics of equations (1) and (2). Equation (2) is not strictly additive so technological patterns don’t follow a discernible pattern. As Proudman and Redding (2000, p. 377) observe, “on the one hand, the presence of section specific learning-by-doing means that initial patterns of international specialization will tend to be reinforced over time. On the other hand, technological transfer and differences in the exogenous rates of productivity growth across sections may both be responsible for reversing initial patterns of international specialization – depending upon the correlation between initial levels of relative productivity and the steady state levels”. To model it empirically the authors consider the evolution of RCA across sections. The empirical model is straightforward and very efficient. The authors assume that the distribution of RCA over time follows a stochastic difference equation (Proudman and Redding, 2000, p. 378). If the space of possible RCA values is divided into m discrete cells, the operator P* (Proudman and Redding, 2000, p. 387) on the stochastic equation becomes a m x m matrix of transition probabilities: (3) Ot P * Ot 1 Where P* contains elements pkl, each of which denotes the probability that an industry moves from cell k to cell l (where k,l ^1,. . . , m`) and which may be estimated by counting the number of transitions out of and into each cell. This estimation already yield interesting results, but the authors go further by estimating formal indices of mobility. For our purpose the transition probabilities present enough information on the patterns of trade between Brazil and the EU. For comparison the benchmark used is that of Proudman and Redding (2000), who provide a pooled estimation of five countries. The pooled estimation gives the transition probabilities of a section moving from a lower state of RCA to a higher one. Moving up has the effect that an economy is gaining RCA on that section, while moving down has the opposite effect. It should be
140
Evolving Patterns of Trade between Brazil and the EU
noted that the results of the analysis are an average probability of the entire trade pattern, i.e., the average probability that a section is going to gain or lose RCA. Following Brasili et al. (2000), the statistical distribution of the results is analyzed, using its marginal density and scatter diagrams to verify the robustness of the results. The econometric analysis complements the statistical analysis in developing a test for the marginal intra-industry trade index (MIIT) (Brülhart, 1994). Bastos and Cabral (2007) analyze the MIIT to describe the dynamics of trade patterns in 20 OECD countries. The MIIT is also used as the dependent variable in econometric analysis (Bastos and Cabral, 2007). We propose an alternative way of using the MIIT to consider the dynamics of trade patterns, by looking at the evolution of the MIIT over time, estimated econometrically, in the same vein as research on terms of trade advanced by Cuddington and Urzua (1989) and Zanias (2005). When applied to terms of trade analysis the idea is to test for an improvement or deterioration of the terms of trade by laying an exponential trend line through the data (Cuddington and Urzua, 1989) or by looking for structural breaks (Zanias, 2005). We apply this test in the context of panel data, and focus on the MIIT. In this way a deterioration or improvement would indicate the deterioration of trade specialization or gains of trade specialization in a sector. Cuddington and Urzua (1989) estimate: ly (t ) a bTIMEit eit (4) If b>0 there is an improvement of the terms of trade or trade patterns, and if b<0 a deterioration. The null hypothesis is that b = 0, i.e. that the pattern of trade is invariant over the period considered, while the alternative is that there is a trend over time. We introduce panel dynamics to incorporate the effect of the evolution of trade patterns in each HS chapter. In this way, accepting the null is stronger in the sense that it would mean that the patterns of trade are invariant over time and over export sectors. Introducing fixedeffects equation (4) then becomes: ly (t ) a bTIME e(t ) (5) Zanias (2005) observes structural breaks in terms of trade series. In this sense, it is relevant to consider the possibility of structural breaks in patterns of trade. A break would indicate that patterns of trade are evolving over time, either due to a shock or to internal dynamics. However, instead of considering ad hoc possible structural break points, our investigation follows the work of Andrews (2003),
Rodrigo M. Zeidan
141
who presents a test for structural break in panel data which is robust to non-normal, heteroskedastic and serially correlated errors, and allows for the number of post break observations to be small (Chan et al., 2006). The test has already been used with trade data (Mancini-Griffoli and Pauwels, 2006), although in a different context. Our idea is to test if there are structural breaks at all. Since the transposition of Andrews (2003) theoretical work to real world data is not direct, we follow Mancini-Griffoli and Pauwels (2006). The regression that is the basis for the test is (Mancini-Griffoli and Pauwels, 2006, p. 4):
Yit
° X it' E 0 U it ® ' °¯ X it E1t U it
1,..., T ½° ¾ T 1,..., T m°¿ t
t
(6)
for individuals i = 1, . . . , n, and where T is the postulated break date. The test naturally hinges on the following hypotheses: H0 : 1t = 0 against HA : 1t 0 The coefficients are homogeneous across i under both the null and the alternative hypothesis. The alternative hypothesis requires all individuals to exhibit an end-of-sample break. The test statistics and the critical values are calculated in the same form as in Mancini-Griffoli and Pauwels (2006), where one can also find the translation of the test to panel data. III.
Brazil and the EU – Some Trade Facts
A. Data sources and data construction The first best data on international trade are those of national government agencies. However, most often than not, the micro data necessary for whole merchandise trade analysis is either unavailable due to bureaucracy or incompatible with other data sets. The second best alternative is to rely on global data from multilateral agencies, like the WTO, World Bank, UNCTAD etc. Using data from these sources (now integrated at the WITS project) yields the benefit of a single compatible data source, which we prefer in this paper. Another issue is which European Union to consider. Since the EU is a free trade zone, it is better to consider the whole EU instead of the Euro-Zone. Also, since the present analysis uses data from 1994 to 2005, it encompasses changes in the EU structure. Those changes refer to the accession of ten countries in 2004, which the trade data should reflect that. However, for consistency, the data
142
Evolving Patterns of Trade between Brazil and the EU
will be backtracked and the European Union will be considered as having 25 countries for the entire data period. Romania and Bulgaria are not considered, but the choice between 15, 25 or 27 countries would not likely affect trade patterns, since most of the trade between Brazil and the European Union is with the large European countries – France, UK, Germany, and the accession of ten countries in 2004 only enlarged trade between Brazil and the EU by four per cent. Data is annual for the 1994-2005 period, both for the European Union and Brazil. The data comprises the whole trade flow disaggregated by 06-digit on the Harmonized System (HS06). A typology of goods divided by technological content and input specialization is also used, sourced from OECD. The result data is then a panel with bilateral flow for each sector for each year and each HS02 chapter, comprising a panel with 97 chapters and twelve years. B. Brazil - EU trade – aggregated data analysis In 1994 Plano Real was implemented in Brazil and inflation was finally curbed down from an average of 2000 per cent in previous years. The early years of Plano Real were marked by an adjustment process to reach low inflation, with significant GDP growth. This soon changed and Brazil experienced a recession period from 1997 to 1999, with a negative impact on Brazil’s trade flows with the world, worsened by an appreciated pegged currency. In the 19941999 period, exports grew by only 1.97 per cent annually, while imports grew by 8.3 per cent per year. Unlike in 1994, for the entire period following Brazil’s trade balance was negative. In contrast to the 1994-1999 period, Brazilian trade flows experienced an impressive growth in the 2000-2005 period, with exports increasing by more than 100 per cent - from U$55 billion in 2000 to U$118 billion in 2005 –, and a cumulative imports growth of 56 per cent (from U$47 billion in 2000 to U$73 billion in 2005). In the same period, the cumulative growth of world trade amounted to 60 per cent. This impressive performance of Brazilian exports is usually attributed to two devaluations of the Real, one in January 1999 - when the government was unable to maintain the pegged exchange rate (the Brazilian Real jumped from U$1-R$1,22 to U$1-R$2), the other in 2002, an election year when the exchange rate went up to almost R$4 for a dollar. Complementary to these devaluations, there are further explanations for the high Brazilian
Rodrigo M. Zeidan
143
trade growth: a higher linkage to world trade growth, a rise in commodities prices, and a change in industrial patterns – a kind of export effort by firms to meet a growing world demand for Brazilian goods. The argument of price changes – the terms of trade explanation - is commonly favored, although Puga (2006) shows that, in fact, Brazilian terms of trade actually worsened in 2005 compared to 2000, even though exports prices are higher, as can be seen in Figure 1, which presents the export and import price index in the 2000-05 period. Figure 1: Brazilian Terms of Trade – Development of Exports and Imports Price Index 2000-05 130
TOT
120 110 100 90 2000
2001
2002 imports
2003
2004
2005
exports
Source: Puga (2006).
Puga (2006) also decomposes the trade flows growth in the period and concludes that only U$11 billion of exports growth can be explained by rising international prices, while the other U$52 billion in growth is not derived from price fluctuations, but from firms’ export efforts to meet an international growing demand. Moreover, at the same time that export prices were growing (in the figure, from the lowest value in 2002 onwards), the exchange rate started to appreciate sharply. Figure 2 shows that the Brazilian currency peaked in October 2002, reaching above 3.8 for both the dollar and the euro. Since then the real has sharply appreciated, while, for the period of July 2002 to early 2005, the real was fluctuating between R$3.3 and R$3.8 per euro. In the same period European trade experienced growth, but in line with world trade growth, at a pace of 8-10 per cent per year. The EU accounts for 13 per cent of world trade. It shows a diversified export and import structure (in some product areas where the EU has no comparative advantages, such as agri-
144
Evolving Patterns of Trade between Brazil and the EU
cultural products for example, it becomes competitive through export subsidies), and is one of the most important trading partner of Brazil, alongside with NAFTA and MERCOSUR. Figure 2: Nominal Exchange Rate – Real, Euro and Dollar 2000-05 4.00 3.50 3.00 2.50 2.00 1.50
Real/U$
Euro/U$
2007-01-01 01/2007
2006-01-01 01/2006
2005-01-01 01/2005
2004-01-01 01/2004
2003-01-01 01/2003
2002-01-01 01/2002
2001-01-01 01/2001
01/1999 1999-01-01
0.50
01/2000 2000-01-01
1.00
Real/Euro
Source: Central Bank of Brazil (2007).
EU exports to Brazil grew at 4.9 per cent annually during 19942005, which was lower than the total Brazilian imports, 6.9 per cent per year. Although this would imply that the EU has become a less important trading partner, this difference can be explained by the high growth of Chinese exports to Brazil, which went up by 25 per cent per year in the same period. Imports from other trading partners of Brazil showed similar growth as the EU, NAFTA five per cent per year, MERCOSUR 3.,2 per cent, and Japan four per cent. Figure 3 shows that the European Union is still the largest exporter to Brazil, even if China is an increasingly relevant partner. The decline of the EU share in Brazilian imports cannot be ascribed simply to exchange rate fluctuations, since NAFTA, whose currency depreciated more against the real, shows a larger decline of its share in the Brazilian market. The fact is that while Brazilian imports are growing, the relative share of EU exports is declining, even if absolute trade flows were growing.
Rodrigo M. Zeidan
145
35 30 25 20 15 10 5 0
China EU Mercosur
20 04
20 02
20 00
19 98
19 96
NAFT A
19 94
per cent of total imports
Figure 3: Brazilian Import Structure by Country and Regions 19942005
Source: WITS (2007), elaborated by the author.
Gains in trade are not only a result of efficiency, but also are, as previously observed, a consequence of technology transfer to recipient countries. To analyze possible gains in trade one has to decompose EU exports to Brazil in terms of their technological content. This decomposition is achieved by using the OECD classification of HS04 products by technological content, and export data from WITS. High and medium-high technological content exports are presented in Figures 4 and 5 below.
60 50
China
40
EU
30
Mercosur
20
NAFT A
10 0 1 99
4 19 95 19 96 19 97 19 98 19 99 20 00 2 00 1 20 02 2 00 3 20 04 2 00 5
per cent of total imports
Figure 4: Brazilian Imports of High-Technology Goods 19942005, by origin
Source: WITS (2007), elaborated by the author.
Figures 4 and 5 show the importance of EU exports in high and medium-high technology goods in the Brazilian market. For both kinds of goods the EU is the leading exporter, although losing
146
Evolving Patterns of Trade between Brazil and the EU
market-share. This is, unsurprisingly, a result of Chinese competition. In the case of high-technology goods, the Chinese competition is very recent, growing only in the last three years of the 1994-2005 period. In contrast, Chinese exports of medium-high technology goods have been growing since the end of the 1990’s. Chinese exports are usually not associated with high-technological content. However, their growth to the Brazilian market shows the diversity of China’s trade expansion. Nevertheless the EU´s relative share of exports of high and medium-high technology goods should stabilize, since the EU is a traditional trade partner and, as will be seen, has comparative advantages in high-tech industries. The EU exports to Brazil are concentrated in medium and hightechnology goods. Brazilian exports to the EU are mostly concentrated in low technology goods, as can be seen in Figure 6.
50 40
China
30
EU
20
Mercosur
10
NAFT A
0 1 99 4 1 99 5 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05
per cent of total imports
Figure 5: Brazilian Imports of Medium-High Technology Goods 1994-2005, by origin
Source: WITS (2007), elaborated by the author.
Two main observations can be derived from the figure above. First, the share of Brazil in the EU market is constant over time, comprising two per cent of EU imports. This is surprising, since Brazilian exports performance was much worse in the 1994-1999 than in later years. The other important point is that Brazilian exports are concentrated in low-technology goods, with the share of these products averaging four per cent of EU imports, in contrast to 1.3 per cent with higher-technology goods. This pattern can be clearly seen in Figure 7, which presents Brazilian exports to the EU decomposed by production factor.
Rodrigo M. Zeidan
147
5.0 4.0 3.0
High Medium-high Medium-low
2.0 1.0
Low T otal
0.0 19 94 19 95 19 96 19 97 19 98 19 99 20 00 2 00 1 2 00 2 2 00 3 20 04 20 05
per cent of total imports
Figure 6: Share of imports from Brazil in EU´s total imports 1994-2005, by technological content
Source: WITS (2007), elaborated by the author.
Figure 7: Share of imports from Brazil in total EU imports – 19942005, by main production factor 12
per cent
10 8 6
Capital Machinery
4
Labor
2
Natural Res.
19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05
0
Source: WITS (2007), elaborated by the author.
Figure 7 reveals a structure that one would expect in NorthSouth trade. The South exports, i.e. Brazil´s exports, are heavily concentrated on natural resources goods, followed by labor-intensive goods. While the EU´s total import share from Brazil is constant, the share of natural resources intensive goods is increasing, reaching almost 10 per cent of EU imports in 2005. This is surprising, because although Brazil is an important exporter of natural resources goods, it is not the main European trading partner, and furthermore Brazil´s exports to the rest of the world are not only growing in natural resources, but in other areas as well. We can
148
Evolving Patterns of Trade between Brazil and the EU
conclude that the Brazil – EU trade pattern represents the common textbook story of North-South trade. But is that the whole story? Looking into different sectors should bring more information regarding trade patterns. C. Brazil - EU trade – disaggregated data analysis by product sector To examine the development of trade patterns, a look on sectoral trade data is important since trade patterns partly reflect product specialization in the Heckscher-Ohlin tradition, and the rest intraindustry trade. While intra-industry trade is most relevant in NorthNorth trade, it is almost disregarded in North-South analysis. However, the industrialization process experienced by developing countries has brought also an advance in intra-industry trade between developing and developed countries. Nevertheless specialization in natural resources and labor intensive goods is still the main characteristic of North-South trade. There is no specific theory on North-South trade, but many works deal with the causes and consequences of trade between developed and developing countries. The main finding is that factor endowments still drive North-South trade, but intra-industry trade is increasingly important due to the fact that the technological ladder hypothesis (where countries improve technologically with ongoing industrialization) leads developing countries into increased production of manufactured goods, which makes technology a determinant of the dynamics of trade. Also, there is trade diversification at the sectoral level, with differentiated products leading to more diversity in trade flows. Before looking at the data, one has to make a choice regarding the level of disaggregation. There are roughly 5200 products at 06 digits and 1200 at 04 digits. At the chapter-level, data at 02 digits yields 97 chapter, which can be further clustered in 21 chapters. Calculations can be done by looking first at 06 or 04 digits level and then averaging to show the results in the more suitable form of 97 or 21 chapters. We shall follow this approach and presenting the main findings using 21 chapters. When regarding Revealed Comparative Advantages, the results will be presented for all 97 chapters. For our analysis we use the marginal intra-industry index of trade (MIIT), the Herfindahl-Hirschman Index (HHI), and the Revealed Comparative Advantage index (RCA), first described by
Rodrigo M. Zeidan
149
Balassa (1965). The present version of the RCA index was modified by Proudman and Redding (2000), and has the advantage of normalizing Balassa’s RCA to maintain a constant mean, equal to one over time. Since the analysis covers a twelve-year period it is necessary to divide it to visualize the results better. The choice is to divide 19942005 in three intervals according to structural changes in the Brazilian and European economies. Our first period is from 1994 to 1998, since 1999 marks the start of the euro and the first currency depreciation in Brazil. The second period is 1999 to 2001, since the end of 2001 marks the beginning of a mild world recession and 2002 presents another sharp currency depreciation in Brazil. The third period goes from 2002 to 2005, the last year with available data. First we calculate the MIIT for these three periods. For averaging the index by trade chapters, a simple procedure is used. Total trade flows of the entire period are summed up (import plus exports plus each year) for of each of the 1200 HS04 goods, and the relative share of product i in total bilateral trade is determined. This share is then used for averaging of the MIIT index to find the chapter average MIIT index. The resulting index is in the (0,1) interval. Values close to zero indicate inter-industry trade, while for intra-industry trade results are close to the unity. Since HS04 disaggregation is used, values close to one rarely happen, since specialization in differentiated goods would make it difficult for industry exports and imports in the exactly same amount for the entire chapter. The marginal intra-industry trade index is defined as (Bastos and Cabral, 2007):
MIITijt
1
| ' t X ijt ' t M ijt |
| ' t X ijt | | ' t M ijt |
(7)
where tXijt=Xijt f – Xijt 0 and tMijt=Mijt f – Mijt 0. Xijt 0 and Mijt 0 are, respectively, the exports and imports of section I from country j at the beginning of period t, while Xijtf and Mijt f are the respective exports and imports at time f. The results of the MIIT for Brazil - EU trade for the 1994-2005 period are presented in Table 1. Table 1 also indicates the average share of each product chapter in the trade of the entire period. The results corroborate the expected evolution of the trade pattern. Agricultural goods are the main source of inter-industry trade, with chapters 1 (meat), 2
150
Evolving Patterns of Trade between Brazil and the EU
(vegetables), and 4 (food preparations) showing absolute values close to zero that decrease. The relevant industrial chapters all show increasing values, with the exception of the automotive industry (chapter 17). In chapters 16 (machinery), 7 (plastics), and 15 (base metals), an increasing specialization in intra-industry trade becomes evident. Table 1: Marginal Intra-industry Trade Index for Brazil - EU trade and share of total trade, 1995-2005 MIIT Chap.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Description meat, animal products and live animals vegetable products animal or vegetable fats and oils prepared foodstuffs; beverages, spirits, tobacco mineral products products of the chemical or allied industries plastics, rubber and articles thereof raw hides and skins, leather, wood, charcoal, cork and articles pulp, paper and printed material textile and textile articles footwear, headgear, etc. stone, plaster, cement, glass and articles precious stones and metals base metals and articles machinery; electrical equipment vehicles, aircraft, associated equipment electronics arms and ammunition; parts miscellaneous manufactured articles works of art, collectors' pieces
share in total trade*
1995-98 1999-01 2002-05 % 0.094 0.073 0.050 2.7 0.027 0.024 0.019 10.5 0.146 0.197 0.156 0.4 0.032 0.020
0.021 0.016
0.023 11.8 0.046 8.0
0.159 0.316 0.044 0.041 0.088 0.203 0.066 0.378 0.074 0.191 0.283 0.410 0.170 0.372 0.284 0.273
0.142 0.354 0.043 0.048 0.093 0.179 0.058 0.404 0.082 0.230 0.316 0.420 0.173 0.581 0.229 0.504
0.180 10.8 0.405 3.3 0.131 0.6 0.020 2.2 0.135 3.2 0.199 1.5 0.041 0.9 0.389 1.0 0.050 0.7 0.242 6.9 0.383 22.8 0.311 9.1 0.215 2.3 0.426 0.0 0.241 1.2 0.258 0.0
* share of exports plus imports of each chapter in total trade 1995-2005, period average.
With respect to diversification of trade, the issue arises whether there is concentration inside exports and imports chapters. The industrialization process should bring more diversification in industrial products. Also with agricultural products, more diversification
Rodrigo M. Zeidan
151
is expected since product diversification is a clear sign of a healthy agricultural country that does not rely on a single product. Trade diversification happens if there is a decreasing share of leading products in a chapter’s exports. This can be examined with the Herfindahl-Hirschman Index (HHI). It is defined as: n
Hj
¦s
2 i
(8)
i 1
where si is the share of product i in chapter j. The HHI ranges from 1/n to 1, with values close to 1/n indicating an export chapter that is truly diverse, with each product having the same share as the others, whereas values close to one indicate that only one product is responsible for the exports of the entire chapter.4 Although the HHI for international trade is rarely done, it is quite useful to examine trade diversification, especially if a time series is available and if changing patterns are expected. The HHI is presented for each chapter for the three periods, based on the share of HS04 products in a chapter’s exports. A decreasing HHI indicates increasing trade diversification, and an increasing HHI that trade becomes more concentrated in a few products. The values of the HHI for Brazilian exports to the EU are presented in Table 2. The HHI indicates that most exports are concentrated in primary products, with chapters 2 (Vegetables), 4 (Food preparations), and 5 (Mineral products) accounting for a little over 50 per cent of Brazilian exports to the EU in 1994-2005. In the most important product chapters no trade diversification takes place, but the pattern is not quite clear, since for some chapters the HHI is either constant (16 –Machinery), or increasing (4 – Food preparations). For most chapters the absolute value of the index is close to zero, indicating an already diversified trade pattern. With respect to EU exports to Brazil, the HHI is given in Table 3 below. European exports to Brazil are concentrated in industrial products accounting for more than 90 per cent of total EU exports to Brazil. Finally, we analyze the Brazil – EU trade pattern by looking at Balassa’s Revealed Comparative Advantage index as modified by 4
There is a normalized version of the HHI that gives an interval of (0,1), but since the interest is in the evolution of the index instead of its absolute value, we use the simple version. Low values indicate a truly diversified trade.
152
Evolving Patterns of Trade between Brazil and the EU
Proudman and Redding (2000). For every country exports i of chapter j Balassa defines RCAij as: Z ij / i Z ij RCAij (9) / Z Z ij ij j i j
¦
¦ ¦¦
Table 2: Diversification of Brazil´s exports to the EU: HHI of product chapters, 1995 – 2005 share in total exports to EU*
HHI Ch. description 1 meat, animal products and live animals 2 vegetable products 3 animal or vegetable fats and oils 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
% ‘95-98 ‘99-01 ‘02-05 0.242 0.206 0.173 4.4 0.431 0.408 0.403 18.7 0.251 0.220 0.174 0.3
prepared foodstuffs; beverages, spirits, tobacco mineral products products of the chemical or allied industries plastics, rubber and articles thereof raw hides and skins, leather, wood, charcoal, cork and articles pulp, paper and printed material textile and textile articles footwear, headgear, etc. stone, plaster, cement, glass, articles precious stones and metals base metals and articles machinery; electrical equipment vehicles, aircraft, associated equipment electronics arms and ammunition; parts miscellaneous manufactured articles works of art, collectors' pieces
0.322 0.324 0.332
20.5
0.426 0.031 0.075 0.159 0.119 0.476 0.078 0.368 0.094 0.260 0.125 0.046 0.174 0.043 0.215 0.270 0.522
13.5 3.1 1.4 1.1 4.0 4.6 1.5 1.6 0.7 1.0 7.5 8.2 5.8 0.5 0.0 1.5 0.0
0.359 0.039 0.072 0.198 0.124 0.685 0.082 0.308 0.063 0.176 0.111 0.043 0.277 0.043 0.285 0.295 0.227
0.326 0.036 0.060 0.272 0.128 0.484 0.065 0.250 0.064 0.230 0.056 0.046 0.153 0.052 0.413 0.293 0.492
* period average 1995-2005.
Proudman and Redding (2000) normalize the RCAij so that the average equals one:
RCAij
RCA ij /
1 N
¦
j
RCA ij
(10)
Rodrigo M. Zeidan
153
Taking Brazil as the exporter, the RCA index of a specific export chapter is calculated by comparing the exports of the chapter to the EU with all exports to the European market, and comparing it to Brazilian exports of this chapter to the world with total exports to the world. It is a measure of relative competitiveness of a country´s exports. Instead of showing the data with only 21 chapters, we present the RCA for the whole HS02 since the data relates export chapters to total exports, information would be lost with only 21 chapters. In fact, a thorough analysis of the evolution of patterns of trade should be done by analyzing the 5200 HS06 products, but that is beyond the scope of the present paper. The results are presented in Table 4. Table 3: Diversification of EU´s exports to Brazil: HHI of product chapters, 1995 – 2005
Ch.
description
1 meat, live animals- and products 2 vegetable products 3 animal or vegetable fats and oils prepared foodstuffs; beverages, spirits, 4 tobacco 5 mineral products products of the chemical or allied 6 industries 7 plastics, rubber and articles thereof 8 raw hides and skins, leather, 9 wood, charcoal, cork and articles 10 pulp, paper and printed material 11 textile and textile articles 12 footwear, headgear, etc. 13 stone, plaster, cement, glass, articles 14 precious stones and metals 15 base metals and articles of base metal 16 machinery; electrical equipment 17 vehicles, aircraft, associated equipment 18 electronics 19 arms and ammunition; parts 20 miscellaneous manufactured articles 21 works of art, collectors' pieces
* period average 1995-2005.
HHI
share in total exports to Brazil*
‘95-98 ‘99-01 ‘02-05
%
0.148 0.123 0.137 0.221 0.131 0.170 0.564 0.443 0.333
0.8 1.2 0.5
0.085 0.092 0.102
2.0
0.115 0.119 0.322
1.8
0.012 0.017 0.019
19.5
0.051 0.156 0.075 0.106 0.030 0.113 0.030 0.211 0.020 0.012 0.221 0.039 0.334 0.077 0.667
5.6 0.1 0.1 1.6 1.6 0.1 1.3 0.3 6.2 39.3 12.8 4.4 0.0 0.8 0.0
0.022 0.121 0.070 0.108 0.017 0.111 0.029 0.234 0.016 0.009 0.121 0.047 0.303 0.183 0.489
0.024 0.106 0.098 0.057 0.019 0.095 0.033 0.155 0.014 0.008 0.135 0.057 0.219 0.250 0.649
HS 02 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Live animals Meat Fish and crustaceans Dairy produce; eggs; honey; other products Products of animal origin Live trees and plants; cut flowers and ornamental Edible vegetables and certain roots and tubers Edible fruit and nuts; peel of citrus fruit or melons Coffee, tea, mate and spices Cereals Products of the milling industry; malt; starches Oil seeds and oleaginous fruits; misc grains Lac; gums, resins and other saps and extracts Vegetable plaiting materials; other products Animal or vegetable fats and oils Preparations of meat, of fish etc Sugars and sugar confectionery Cocoa and cocoa preparations Preparations of cereals, flour, starch or milk Preparations of vegetables, fruit, nuts or others Miscellaneous edible preparations Beverages, spirits and vinegar Soy beans, Prepared animal feed, residues Tobacco and manufactured tobacco substitutes
Description
Exports EU to BR 95/98 99/01 01/05 0.67 1.03 0.42 0.05 0.05 0.02 0.39 0.40 0.77 1.84 0.63 0.22 2.48 4.05 4.42 0.10 0.17 0.23 0.87 1.08 0.40 0.98 1.09 0.98 0.21 0.37 0.34 0.31 0.10 0.91 4.46 3.15 3.47 0.50 0.93 1.16 1.74 1.97 2.34 0.16 0.24 0.33 2.10 2.15 2.54 0.39 0.26 0.10 0.48 0.39 0.25 0.71 0.29 0.24 0.75 0.63 0.33 1.51 1.18 0.68 0.83 0.84 0.56 0.95 0.57 0.48 0.56 0.74 1.68 0.11 0.18 0.15 0.0 0.0 0.1 0.4 0.2 0.0 0.1 0.2 0.0 0.1 0.6 0.1 0.1 0.0 0.5 0.0 0.1 0.2 0.1 0.2 0.2 0.9 0.1 0.1
%
Exports BR to EU 95/98 99/01 01/05 0.00 0.00 0.45 1.88 1.92 1.34 0.44 1.11 2.02 0.08 0.25 1.23 1.68 1.76 1.79 2.73 3.24 2.79 1.13 0.42 1.60 1.85 2.24 2.60 1.93 2.44 2.39 0.01 0.66 1.15 0.25 0.56 0.81 2.58 2.86 1.96 1.86 1.98 1.25 0.80 0.38 0.63 0.42 0.40 0.35 2.22 2.59 2.81 0.15 0.11 0.14 0.64 0.57 0.41 0.04 0.05 0.16 2.06 2.53 2.23 0.44 0.54 0.83 0.68 0.70 0.59 2.41 3.57 3.34 1.44 1.81 1.18
Table 4: RCA – EU exports to Brazil and share of product group in total trade, 1995 – 2005 (to be continued)
0.0 3.4 0.5 0.0 0.4 0.1 0.0 1.6 6.7 0.3 0.0 10.0 0.1 0.0 0.3 1.9 0.2 0.2 0.0 4.5 0.3 0.2 11.0 2.4
%
HS 02 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
Salt; sulfur; stone; plastering materials, cement Ores, slag and ash Oil, Mineral fuels, and distillation Inorganic chemicals Organic chemicals Pharmaceutical products Fertilizers Tanning or dyeing extracts; dyes, pigments, paints Essential oils and resinoids; perfumery, cosmetic Soap, organic surface-active agents, Albuminoidal; modified starches; glues; enzymes Explosives; pyrotechnic products; matches Photographic or cinematographic goods Miscellaneous chemical products Plastics and articles thereof Rubber and articles thereof Raw hides and skins (not fur skins) and leather Articles of leather Fur skins and artificial fur; manufactures thereof Wood and articles of wood; wood charcoal Cork and articles of cork Manufactures of straw, of esparto, other Pulp of wood or of other fibrous cellulosic material Paper and paperboard; articles of paper pulp
Description
Exports EU to BR 95/98 99/01 01/05 0.69 0.65 0.82 2.42 1.53 1.56 1.07 0.93 0.85 1.41 1.74 2.27 2.45 2.07 1.68 1.12 1.33 0.97 4.59 9.18 13.6 1.77 1.94 1.81 0.65 0.76 0.57 1.84 1.60 1.61 1.09 1.11 1.22 0.40 0.51 0.50 1.10 1.65 1.75 1.51 1.93 2.83 1.45 1.54 1.69 2.15 2.04 2.56 0.78 0.87 0.62 0.08 0.12 0.10 0.01 0.01 0.01 0.09 0.17 0.06 0.84 0.61 0.50 0.74 1.16 0.45 0.05 0.20 0.34 0.95 0.65 1.05 0.2 0.1 1.5 1.2 7.1 4.2 1.6 1.3 0.6 0.5 0.3 0.0 0.4 2.3 3.9 1.6 0.1 0.0 0.0 0.1 0.0 0.0 0.0 1.0
%
Exports BR to EU 95/98 99/01 01/05 2.28 2.64 2.02 1.50 1.77 1.33 1.15 0.87 0.63 0.56 1.00 0.92 0.80 0.78 0.79 0.62 0.24 0.65 0.04 0.04 0.03 1.03 1.02 0.44 0.77 0.89 0.82 0.43 0.21 0.32 1.08 1.26 1.32 0.54 0.45 0.18 0.25 0.23 0.21 0.45 0.69 0.84 0.35 0.48 0.61 0.64 0.68 0.72 1.33 1.66 1.61 1.74 1.73 1.25 2.17 1.64 2.16 1.70 1.70 1.26 0.15 0.22 0.45 2.77 1.30 1.76 1.35 1.68 1.83 0.83 0.36 0.95
Table 4 (cont.): RCA – EU exports to Brazil and share of product group in total trade, 1995 – 2005 1.7 10.5 1.3 0.7 1.2 0.2 0.0 0.2 0.2 0.0 0.2 0.0 0.1 0.3 0.7 0.7 1.0 0.1 0.0 4.0 0.0 0.0 3.8 0.7
%
HS 02 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 68 69 70 71 72 73
Printed books, newspapers, etc Silk Wool, fine or coarse animal hair Cotton Other vegetable textile fibers Man-made filaments Man-made staple fibers Wadding, felt and nonwovens; special yarns Carpets and other textile floor coverings Special woven fabrics; tapestries Impregnated, coated, or laminated textile fabrics Knitted or crocheted fabrics Articles and clothing access, knitted or crochet Articles and clothing access, not knitted or crochet Other made up textile articles Footwear, gaiters and the like; parts Headgear and parts thereof Umbrellas, walking sticks, seatsticks, whips Articles of stone, plaster, cement, asbestos Ceramic products Glass and glassware Precious or semi-precious stones,precious metals Iron and steel Articles of iron or steel
Description
Exports EU to BR 95/98 99/01 01/05 1.23 0.82 0.54 0.10 0.16 0.19 0.08 0.13 0.07 0.64 0.45 0.15 0.48 0.89 0.26 1.28 1.37 1.37 0.40 0.88 0.90 1.35 1.08 0.87 0.61 0.50 0.24 0.54 0.53 0.38 1.87 1.57 1.42 0.62 0.40 0.14 0.10 0.22 0.11 0.06 0.18 0.15 0.18 0.17 0.16 0.08 0.09 0.09 0.28 0.26 0.29 0.37 0.45 0.16 1.13 1.15 1.22 0.55 0.45 0.50 0.89 1.05 1.15 0.12 0.21 0.18 0.73 0.80 0.87 0.93 1.24 1.40 0.6 0.0 0.0 0.1 0.1 0.4 0.2 0.2 0.1 0.1 0.3 0.0 0.1 0.1 0.0 0.0 0.0 0.0 0.5 0.3 0.5 0.3 1.4 1.9
%
Exports BR to EU 95/98 99/01 01/05 1.49 1.74 1.61 0.59 0.59 0.63 1.50 2.41 2.02 0.48 0.68 0.73 2.10 2.33 1.24 0.73 0.30 0.30 0.42 0.19 0.20 0.34 0.32 0.23 1.12 1.47 0.97 1.03 0.89 0.29 0.47 0.34 0.16 0.01 0.05 0.26 1.27 0.61 1.17 0.65 0.34 0.86 1.43 0.92 0.65 0.76 0.58 0.78 0.98 0.60 0.75 0.48 0.13 0.56 0.66 0.66 0.57 0.72 0.43 0.39 0.67 0.89 0.99 1.13 0.81 1.06 0.63 0.79 0.65 0.40 0.36 0.41
Table 4 (cont.): RCA – EU exports to Brazil and share of product group in total trade, 1995 – 2005
0.1 0.1 0.1 0.3 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.2 0.1 0.4 1.6 0.0 0.0 0.3 0.2 0.2 1.0 3.7 0.3
%
HS 02 74 75 76 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
Copper and articles thereof Nickel and articles thereof Aluminum and articles thereof Lead and articles thereof Zinc and articles thereof Tin and articles thereof Other base metals; cermets; articles thereof Tools, implements, cutlery, spoons and forks Miscellaneous articles of base metal Machinery and mechanical appliances Electrical machinery and equipment and parts Railway or tramway locomotives, and parts Vehicles and parts and accessories thereof Aircraft, spacecraft, and parts thereof Ships, boats and floating structures Electronics Clocks and watches and parts thereof Musical instruments; parts and accessories Arms and ammunition; parts and accessories Furniture; bedding, mattresses, mattress supports Toys, games and sports requisites Miscellaneous manufactured articles Works of art, collectors' pieces and antiques
Description
Exports EU to BR 95/98 99/01 01/05 0.55 0.63 0.73 0.98 1.13 2.36 1.01 1.76 1.81 0.75 0.60 0.68 0.64 0.27 0.47 0.46 0.26 0.40 2.12 2.41 2.18 2.04 1.58 1.19 1.74 1.37 1.63 1.89 1.73 1.48 1.51 1.74 1.27 0.51 0.68 0.32 1.64 0.98 0.79 0.70 1.83 1.59 0.07 0.20 0.08 1.46 1.20 1.18 0.51 0.38 0.42 0.50 0.22 0.23 0.14 0.45 0.11 0.31 0.57 0.49 0.26 0.23 0.11 0.94 0.95 0.72 0.07 0.02 0.03 0.3 0.1 1.0 0.0 0.0 0.0 0.1 0.7 0.6 27.1 12.3 0.3 9.9 2.6 0.0 4.3 0.0 0.0 0.0 0.6 0.1 0.2 0.0
%
Exports BR to EU 95/98 99/01 01/05 0.91 0.76 0.35 0.40 1.91 1.55 1.06 1.48 1.15 0.51 0.85 1.03 0.51 0.24 0.42 0.77 0.14 0.84 1.35 1.03 1.83 0.59 0.72 0.73 0.53 0.49 0.53 0.79 1.00 0.96 0.64 0.50 0.69 0.46 0.24 0.23 0.61 0.65 0.29 1.41 1.49 0.64 1.20 0.27 0.45 0.92 0.84 1.16 1.48 1.85 1.36 1.69 1.16 1.35 0.42 0.35 0.25 1.98 1.94 1.52 0.34 0.12 0.54 0.40 0.27 0.47 1.71 0.98 1.79
Table 4 (cont.): RCA – EU exports to Brazil and share of product group in total trade, 1995 – 2005 0.2 0.2 2.6 0.0 0.0 0.1 0.0 0.2 0.1 6.2 2.0 0.0 3.1 2.7 0.0 0.4 0.0 0.0 0.0 1.4 0.0 0.1 0.0
%
158
Rodrigo M. Zeidan
The RCA reveals important information. First, it provides further corroboration that Brazil-EU trade is mostly inter-industry trade, since most Brazilian exports with high RCA are concentrated in agricultural products, whereas industrial goods show a high RCA in EU exports. Moreover, the RCA values are fairly constant over the three periods considered. It does not seem that there is increasing specialization, but that specialization is constant. We shall more formally test this later, following Proudman and Redding (2000). However, the rapid growth of the bilateral trade in the period analyzed, especially in 2000-2005, should have brought changes in the pattern of trade. Since Brazilian trade with the world underwent an increase of almost 100 per cent in 2002-2005, one would expect that this had been caused by an introduction of new products and trade diversification. While, as shown, trade diversification has indeed taken place in the last period, trade specialization still dominated, and Brazilian exports maintained its RCA pattern in the European market. Considering EU exports to Brazil, in all product chapters with high RCA, this remains over one for the entire period, with the exception of chapter 30 (pharmaceutical products), which presents a RCA of 0.97 in the last period. However, in all those chapters the RCA has decreased, which can be interpreted as a loss in competitiveness of EU exports in the Brazilian market, probably due to increased competition form China, as previously observed, or other unobserved factors. In a Heckscher–Ohlin thinking, the conclusion is that trade between countries with similar factor endowments would not occur due to specialization in the same industries. But including product differentiation and economies of scale the implication is that specialization in different products inside an industry is viable and thus intra-industry trade emerges. That means that one of the main conditions for intra-industry trade is that both countries’ industries present RCA above average. Regarding the Brazil - EU trade, this only appears in six of the 97 HS02 chapters. These six chapters account for less than seven per cent of the bilateral trade. Although intra-industry trade between EU and Brazil is growing, the fact that the trade is highly specialized, and that there is no congruence for RCA over one in most industries limits the possibility of intra-industry growth. The growth of Brazilian exports, then, cannot be explained by a rapid industrial development, and the export effort (Puga, 2006) by Brazilian firms seems to be a consequence of growth in the same chapters that already present above average
Evolving Patterns of Trade between Brazil and the EU
159
RCA, with specialization based on labor and natural resources abundance. Thus it seems that technology transfer is not playing a major role in the dynamics of Brazil-EU trade. If so, RCA would have changed over time, and intra-industry trade would have grown at a higher rate. Also, RCA index for Brazilian exports would have grown across the board for industrial products, not only for chapters 90 (Electronics), 72 (Lead), 82 (Tools), and 35 (Albuminoids). IV.
Econometric results
The results of the analysis of the mobility or persistence of trade patterns investigating transition probabilities are presented in Table 5. The interpretation of the table is as follows. RCAs are divided into four sets, numbered 1 to 4, from lower RCAs levels to higher RCAs levels. Upper endpoints are determined at which point transition from one set to another happens. A transition to a higher RCA means a jump from a lower state to a higher state. It can also happen that an export section loses RCA and jumps back to a lower state. The first column gives the upper endpoint in parentheses for each sample, i.e., the cutting point at which RCA transition happens. The numbers on the diagonal are the probabilities that the RCA is constant. Numbers to the left of the diagonal are probabilities that a section is moving up the RCA ladder and numbers to the right indicate a downward movement. The results are iterated to present a five-year average. Table 5: Five-year transition probabilities for Brazil and EU trade Upper endpoint Brazil exports to the EU
EU exports to Brazil
(0.45) (0.86) (1.12) f (0.53) (0.91) (1.24) f
1 2 3 4 1 2 3 4
1
2
3
4
0.7813 0.2103 0.0428 0.0001 0.7936 0.2756 0.0218 0.0041
0.1712 0.5826 0.2032 0.0255 0.1830 0.5109 0.1208 0.0243
0.0311 0.1879 0.6643 0.2566 0.0377 0.2094 0.7767 0.1398
0.0005 0.0155 0.1077 0.7322 0.0092 0.0221 0.0875 0.8188
160
Rodrigo M. Zeidan
Concerning Brazil’s export to the EU one notes that chapters with very high or very low RCAs have a higher probability of staying in such states, while mobility is higher in chapters with average RCAs indexes. In the case of European exports to Brazil results are ever more pronounced, with a 80 per cent probability that a chapter with a RCA of under 0.53 or over 1.24 stays in the same RCA state in a five-year period. In chapters with intermediate levels of RCA the probability of moving upwards or downwards the RCA ladder is low. This is surprising since the impressive growth of the trade flow between Brazil and the EU might have brought changes in specialization patterns. Our analysis provides no evidence for this and confirms the preliminary conclusions drawn in the last section. The last test is a simplified one based on Mancini-Griffoli and Pauwels (2006), since the idea is not to look deeply into structural breaks but to find one. For that purpose, the test is applied to the last year and then backtracked to explore the possibility of a structural break. The “break date” is defined as the first period when the null of stability occurs with at least 90 per cent confidence, for a post-break sample of six quarters. The break criterion is a year. A dynamic OLS is run, trying to find a structural break. Not even at the 10 per cent probability level one is found (see Table 6). This was expected after the strong evidence already found that the trade pattern between Brazil and the EU has been constant over the period analyzed. Table 6: Structural Break test Year 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
=1% 0.027 0.037 0.065 0.041 0.077 0.071 0.030 0.071 0.088 0.078 0.080
=5% 0.033 0.047 0.024 0.027 0.035 0.024 0.066 0.010 0.008 0.029 0.064
=10% 0.044 0.007 0.009 0.000 0.003 0.032 0.037 0.026 0.024 0.022 0.045
Evolving Patterns of Trade between Brazil and the EU
161
V. Final Comments
The goal of this paper was to discuss current issues in North-South international trade in the context of the trade between Brazil and the European Union. First, a comprehensive exploratory data analysis of the patterns of trade was presented, using tools such as the marginal intra-industry index, the Herfindahl-Hirschman index, and the Revealed Comparative Advantage index. The results of this investigation revealed a typical North-South trade pattern between the EU and Brazil, with specialized trade based on factor endowments, and some first weak signs of increasing intra-industry trade and trade diversification. For verifying these preliminary results two econometric analyses were implemented, one investigating the mobility of trade patterns following Proudman and Redding (2000) and another searching for structural breaks in trade patterns following Chan et al. (2006) and Mancini-Griffoli and Pauwels (2006). The second econometric estimation focused on a new use on the tools developed by the authors, based on the marginal intra-industry index. The results confirmed the initial analysis of patterns of trade between Brazil and EU and corroborated that the trade between them is still based on factor endowments and inter-industry trade, with low probability that export sectors either gain or lose RCA over the years. The period analyzed, 1994-2005, is a relatively long period, during which trade between Brazil and the EU experienced expressive growth. One would expect that in such a long period many changes in trade patterns had happened. However, this conjecture had to be rejected. It can be affirmed that the trade pattern between Brazil and the European Union is almost the quintessential textbook description of North-South trade. This pattern will probably be extended by further trade negotiations, especially if the EU cuts subsidies on agriculture. If this happens, Brazilian exports to the EU should soar, since even with many non-tariff barriers Brazil is already responsible for 10 per cent of EU imports of natural resources based goods. In absence of technological knowledge spillovers and endogenous technological progress trade theory predicts that initial specialization patterns will be locked-in over time (Proudman and Redding, 2000) since factor endowments will be the only source of trade. As uninteresting as it is, the present main conclusion is that this seems to be happening in trade between Brazil and the
162
Rodrigo M. Zeidan
European Union. Despite all interesting models showing the relevance of technological spillovers and the technological ladder hypothesis, the North-South pattern of trade between Brazil and the EU seems indeed locked-in. There are a number of further trade issues remaining to be investigated. An interesting issue arises in the context of the accession of new EU members which may have resulted in trade diversion. Another question relates to trade creation which may result from reduced agricultural export subsidizes of the EU, or a reduction of trade tariffs and non-tariff barriers by Brazil. The present analysis did not focus on tariff effects since after the first trade liberalization of the Brazilian economy in the early 1990s no further movements to open up to trade occurred in Brazil. The EU is more open, but has still a closed agricultural sector which has yet to experience significant international competition. To summarize, this paper strived to advance the literature by showing that, in the case of Brazil and EU trade, patterns of trade are not evolving over time, a result which can be classified as unexpected, and which invites a formal test of a Heckscher-Ohlin model to explain it.
References
Andrews, D. W. K. (2003), End-of-sample Instability Tests, Econometrica 71 (6), 1661-1694. Balassa, B. (1965), Trade Liberalization and 'Revealed' Comparative Advantage, Manchester School 33 (2), 99-123. Bastos, P., Cabral, M. (2007), The Dynamics of International Trade Patterns, Review of World Economics/Weltwirtschaftliches Archiv 143 (3), 391-415. Bergoeing, R., Kehoe, T. J. (2001), Trade theory and trade facts, Staff Report 284, Federal Reserve Bank of Minneapolis. Brasili, A., Epifani, P., Helg, R. (2000), On the Dynamics of Trade Patterns, De Economist 148 (2), 233-58.
Evolving Patterns of Trade between Brazil and the EU
163
Brülhart, M. (1994), Marginal intra-industry trade: Measurement and relevance for the pattern of industrial adjustment, Review of World Economics /Weltwirtschaftliches Archiv 127(3), 600-613. Cameron, G., Proudman, J., Redding, S. (2005), Technological convergence, R&D, trade and productivity growth, European Economic Review 49, 775 – 807. Central Bank of Brazil (2007), www.bcb.gov.br/?TXCAMBIO. Connolly, M. P. (1997), Technological Diffusion through Trade and Imitation, Federal Reserve Bank of New York Staff Reports no. 20. Chan, F., Mancini-Griffoli, T., Pauwels, L.L. (2006), Stability Tests for Heterogeneous Panel Data, HEI Working paper no. 24, Graduate Institute of International Studies, Geneva, Economics Section. Chui, M., Levine, P., Mansoob, M.S., Pearlman, J. (2002), North– South Models of Growth and Trade, Journal of Economic Surveys 16 (2), 123-66. Cuddington, J. T., Urzua, C.M. (1989), Trends and Cycles in the Net Barter Terms of Trade: A New Approach, Economic Journal 99, 426-42. Curtis, J. M., Shenjie, C. (2003), Trade Costs and Changes in Canada's Trade Pattern, World Economy 26 (7), 975-82. Hoppe, M. (2005), Technology Transfer Through Trade, FEEM Working Paper No. 19.05. Hummels, D. (1999), Toward a Geography of Trade Costs, University of Chicago, mimeo. Irwin, D. A. (2002), Long-run trends in world trade and income, World Trade Review 1, 89-100. James, W. E. (1994), Changing patterns of trade in goods and services in the Pacific region: Market-driven economic integration, Business Economics 29 (2), 14-21. Kaplinsky, R., Santos-Paulino, A. (2006), A disaggregated analysis of EU imports: the implications for the study of patterns of trade and technology, Cambridge Journal of Economics 30 (4), 587611.
164
Rodrigo M. Zeidan
Keller, W. (2002), Trade and the Transmission of Technology, Journal of Economic Growth 7 (1), 5-24. Lumenga-Neso, O., M. Olarreaga, M. Schiff, (2001) On ‘indirect’ trade-related research and development spillovers, Policy Research Working Paper Series 2580, The World Bank. Mancini-Griffoli, T., Pauwels, L.L. (2006), Did the euro affect trade? Answers from end-of-sample instability tests, HEI Working Paper no. 04-2006, Graduate Institute of International Studies, Geneva, Economics Section. Proudman, J., Redding, S. (2000), Evolving Patterns of International Trade, Review of International Economics 8(3), 373-396. Puga, F. P. (2006), Porque crescem as exportações brasileiras, BNDES Visão do Desenvolvimento no. 1, http://www.bndes.gov.br/conhecimento/visao/visao_01.pdf (last visit: 08 September 2008). Requena-Silvente, F., Walker, J. (2007), Is Inequality Harmful for North-South Intra-Industry Trade Growth?, International Trade Journal XXI(1), 25-52. Sachs, J., Yang, X., Zhang, D. (2002), Pattern of Trade and Economic Development in a Model of Monopolistic Competition, Review of Development Economics 6 (1), 1-25. Stehrer, R., Wörz, J. (2003), Technological convergence and trade patterns, Review of World Economics / Weltwirtschaftliches Archiv 139 (2), 191-219. Uchida, Y., Cook, P. (2005), The effects of competition on technological and trade competitiveness, The Quarterly Review of Economics and Finance 45 (2-3), 258-283. Wacziarg, R., Welch, K.H. (2003), Trade Liberalization and Growth: New Evidence, NBER Working Paper No. 10152. WITS (2007), World Bank, http://wits.worldbank.org/ World Economic Outlook (2007), IMF, available http://imf.org/external/pubs/ft/weo/2007/01/index.htm.
at:
Yi, K.M. (2003), Can Vertical Specialization Explain the Growth of World Trade?, Journal of Political Economy 111 (2), 52-102. Zaghini, A. (2005), Evolution of trade patterns in the new EU member states, The Economics of Transition 13 (4), 629–658.
Evolving Patterns of Trade between Brazil and the EU
165
Zanias, G. (2005), Testing for Trends in Terms of Trade Between Primary Commodities and Manufactured Goods, Journal of Development Economics 78 (1), 49-59. Zhang, W. B. (1994), Knowledge, growth and patterns of trade, Annals of Regional Science 28 (3), 285-304.
Maria Bas and Ivan Ledezma
Trade Integration in Manufacturing: The Chilean Experience Abstract Using bilateral trade flow data from 1979 to 1999, we estimate trade integration between Chile and its principal trading partners (European Union, United States and Latin America) during this period. Our estimates are based on a gravity specification, theoretically grounded on a monopolistic competition framework with increasing returns. Trade barriers are measured following the border effect methodology by comparing inter-national imports to intranational ones. Our results are consistent with the agenda of trade integration followed by Chile. Moreover, trade integration turns out to be heterogeneous across industries and over the time. We also find asymmetries between export and import oriented policies. All these features are usually missing when using only direct measures of trade policies. I. Introduction Despite common views of world globalization, the assessments about the level of trade integration of an "open" country reveal conflicting results (Anderson and van Wincoop, 2004). The reason is that trade costs involve not only direct policies such as tariffs, nontariff barriers and exchange rates, but also more indirect integration policies like insurance, infrastructure, information networks, enforcement laws and property rights. The picture becomes still more complex if language, preferences and tastes are taken into account to encompass a potential home bias in consumption. In this context, the standard use of tariff data or aggregate trade ratios turns out to capture only poorly well known international examples of trade integration. The Chilean experience is one of these examples. Chilean dictatorship, in power from 1973 to 1990, implemented a deep package of market reforms concerning every economic field. Among them, trade liberalization took place in the second half of
168
Maria Bas and Ivan Ledezma
the 1970s. Since the beginning of the military government, all trade barriers and restrictions to trade were removed. Average nominal tariff decreased gradually from 98 per cent in 1973 to 10 per cent in 1979 (see Figure 1). In particular during the 1990s, Chilean government further achieved its commitment to promote trade by initiating several trade agreements with different countries and regions (see Table 1) including almost all Latin American countries, the United States (US), the European Union (EU) and also Asian countries in recent years. However, if one looks only at nominal tariffs there is only a modest change after the mid 1980s, which in addition is almost homogeneous across industries. Nominal tariffs document the radical unilateral trade liberalization of Chile in the late 1970s, but hide important Chilean policy efforts to boost exports and the international political integration achieved after transition to democracy. On the other hand, aggregate trade flows are a poor measure because of a lack in detail and the bias induced by "exogenous" shocks on productivity and demand. The aim of this paper is to obtain deeper insight into Chile's integration strategy by distinguishing between export and import oriented policies. Using bilateral industry-level trade data of Chile and its main trading partners in manufacturing (European Union, the Unite States and Latin America), we measure the evolution of market access across industries in both direction of trade. Since the strategy of the Chilean government was to set a combination of trade policies, we need to be able to capture the effect of the full set of barriers faced by Chilean exporters to access foreign markets and for Chile´s trade partners to enter it. This also involves identifying potential asymmetries in trade protection and their evolution. The method used in the empirical literature of border effect fits these requirements. This method makes use of gravity equations and provides a measure of international trade integration by considering, as a very intuitive benchmark, the market access of domestic producers to domestic markets (intra-border trade). Consider for instance two countries. If, after crossing the national border, producers face no further barriers to trade than those already present in the national market, both countries will be consider as completely integrated. The pioneering work of McCallum (1995) uses this strategy along with detailed intra-national trade data to study the market access between Canada and the US. Despite their expected high trade inte-
Trade Integration in Manufacturing: The Chilean Experience
169
gration, trade between the US and Canada was found to be about 22 times more difficult than Canadian intra-national trade. Figure 1: Evolution of average nominal tariffs in Chile 1975-2000 0.55 0.45 0.35 0.25 0.15
19 75 19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99
0.05
Nominal Tariffs
Source: Central Bank of Chile.
Such an assessment, of course, should be done keeping everything else equal. Anderson and van Wincoop (2003) demonstrated that the lack of theoretical foundation might lead to a bias in the estimated border effects. Namely, omitted variables concerning multilateral price indexes would have induced an upward bias in the estimates of McCallum (1995), though the "corrected" border effect still remains important regarding direct trade cost measures between both countries. In order to identify a set of determinants of trade that should be controlled for, we follow Fontagné, Mayer and Zignago (2005) and Head and Mayer (2001) who use a specification based on the Dixit and Stiglitz (1977) monopolistic framework with increasing returns as in Krugman (1980). Their theoretically founded gravity equation relies on Wei (1996) who solves the problem of the lack of information on internal national flows by simply considering production minus exports. Using this specification we remain as close as possible to theory and keep a nice feature of the border effect literature: the measure of eventual asymmetries in protection. Since in this paper we focus on disentangling export and import barriers, the rigorous exercise of general equilibrium comparative statics of Anderson and van Wincoop (2003) is less convenient for our purpose as it assumes symmetry in trade costs across countries. Previous studies regarding the trade integration of Chile mainly focused on macro-economic and welfare effects of trade
170
Maria Bas and Ivan Ledezma
agreements using general equilibrium models (Harrison et al. 2003, Chumacero et al.2004, Cabezas 2003). An exception is Herzer et al. (2007) who use a gravity approach and analyze the impact of trade agreements between Chile and the EU on the structure of Chile´s exports predicting an important export increase in food industries (mainly fish, fruits and beverages). Our contribution to this literature consists of providing an assessment of the evolution of trade integration at the industry level considering all barriers to trade. By doing so, we obtain measures of Chile´s trade integration attributing to heterogeneity across sectors and over the time. The rest of the paper is organized as follows. In Section II the estimation strategy is presented. Section III presents the database used in the estimations. Section IV discusses the main results of border effects estimations at national and also at industry level. Section V concludes. II. Methodology and data A. Estimation strategy The empirical literature has proposed several applications of gravity equations. We are particularly interested in those comparing intranational trade flows (within a country) to international trade flow (between countries). As previously mentioned, we follow the methodology of Fontagné et al. (2005) based on the gravity model derived by Head and Mayer (2001). This approach seems to be best suited to measure Chilean trade integration as it corrects for the lack of theoretical foundations of earlier works, keeping the intuitive strategy of using intra-national trade as a benchmark of trade integration. Let us consider a CES utility function in which the (representative) consumer of country i has specific preferences aijh for each variety h depending on the country of origin j (the exporter). V
V 1 º V 1 ª N Mj Ui « aijh cijh V » (1) «¬ j 1 h 1 »¼ Where >1 is the constant elasticity of substitution between domestic and foreign differentiated varieties. Varieties belonging to the same country share the same weight in the utility function. Thus, under the assumption of symmetric producers we can drop the h subscript and consider the whole set of M j varieties. This
¦¦
Trade Integration in Manufacturing: The Chilean Experience
171
standard CES formulation implies the following import expenditure of country j: 1V
§ pij · (2) mij cij pij ¨ ¸ M jYi ¨ aij Pi ¸ © ¹ Imports are valued at the point of consumption (in c.i.f) pij p jW ij and include the producer price (in f.o.b) pij augmented by all transaction costs related to trade, modelled as iceberg N
costs W ij . Total expenditure of country i Yi
¦ mij considers all
j 1
imports, including intra-national ones. This gravity-like equation in (2) summarizes in 1
ª N § p ·1V º 1V ij ¸ « ¨ Pi M j» (3) « ¦ ¨ aij ¸ » j 1 ¹ «¬ © »¼ the consumer prices of all varieties consumed in country i. This index allows us to account for the critique of Anderson and van Wincoop (2003) about the absence of theoretical foundations and the assumption of identical prices across countries. Pi is a multilateral variable that takes into account asymmetries across countries in price setting, namely size, as well as trade barriers and preferences. If omitted, not only a multilateral control is missing but also a bias is generated between the error term and the usual dummies of trade partners. Anderson and van Wincoop (2003) argue that the omission of multilateral price effects, what they call "multilateral resistances", explains the upward bias in border effects of Canada vis-à-vis the US estimated by McCallum (1995). There are four possible approaches to estimate this equation consistently capturing price effects in a theoretical index. First, one can use price index data. Bergstrand (1989) and Baier and Bergstrand (2001) follow this strategy and measure prices with GDP deflators. However, as highlighted by Anderson and van Wincoop (2004), empirical counterparts of Pi such as CPI measures neglect changes in the true set of varieties and do not accurately reflect non tariff barriers and indirect trade policies. The second method is the one proposed by Anderson and van Wincoop (2003). They develop a method using estimated border effects to measure price effects. This estimation strategy is based on
172
Maria Bas and Ivan Ledezma
two steps. The first one consists in estimating the gravity equation using a non-linear least squares estimator to obtain the parameters. In the second step, they remove border barriers and calculate the change in bilateral trade flows to determine the impact of national borders on trade flows. Besides practical difficulties of implementation,1 a crucial limitation for our purposes, as we previously explained, is the assumption of symmetry in bilateral trade costs. The third approach is to use a fixed effects specification to measure unobservable prices. Price indices are considered as unobserved variables and are measured as the coefficients of individual fixed effects related to source and destination (Harrigan 1996, Redding and Venables 2004). Feenstra (2003) shows that the coefficients of fixed effects estimations are consistent and their values are very similar to those obtained by Anderson and van Wincoop (2003). If the economic and geographic determinants captured by fixed effects vary over time, a useful strategy consists in eliminating the price index in equation (2) by expressing inter-national imports mij relative to intra-national ones mii . Head and Mayer (2001) use this approach. Under monopolistic competition and symmetric technology, quantities are symmetric in equilibrium and the number of firms of each country is obtained by simply dividing output value vj v j by firm output value: M j . Thus, equation (2) can be then qp j written as: V 1 V 1V § a ij · § p j · § W ij · §vj · ¨ ¸ ¨ ¸ ¨ ¸ ¨ ¸ (4) ¨ p ¸ ¨W ¸ ¨v ¸ mii ¨© aii ¸¹ © i ¹ © ii ¹ © i ¹ To obtain and empirical counterpart of this gravity equation, Fontagné et al. (2005) state that trade costs W ij are composed of
mij
distance d ij (related to transport costs), ad-valorem tariffs t ij and the "tariff equivalent" of non tariff barriers NTBij .
1
One should implement a routine in which, at each iteration, the sum of squared residuals of the gravity equation is minimized to solve simultaneously the estimation of trade costs and the multilateral price index.
Trade Integration in Manufacturing: The Chilean Experience
173
Wij p j { dij 1 tij 1 NTBij p j
(5)
G
The structure of protection varies across all pairs of partners and depends on the direction of the flow for a given pair. To capture this protection framework, taking the example of the US as trade partner, the following dummy structure is defined:
1 t 1 NTB { exp ª¬K US _ CHL ij
ij
ij
J CHL _ USij º¼
(6)
Where US _ CHLij is a dummy variable set equal to 1 when j (the exporter) is Chile and i (the importer) the US. Similarly, CHL _ US ij is a dummy variable set equal to 1 when j is the US and i Chile. Preferences aij are composed of a random component
eij and a coefficient E i , representing a systematic preference for goods produced in the home country. This "home market bias" is reduced to E i OL OC when the countries share the same
language Lij
1 and are contiguous Cij define preferences as:
[
1 . These assumptions
]
aij { exp eij E i OL Lij OC Cij
US _ CHL
ij
CHL _ USij
(7)
Combining equations (4)-(7) and adding the subscript s for industry-level variables, the following estimable equation is obtained: § mijs · § v js · § dij · ln ¨ ¸ ln ¨ ¸ V 1 G ln ¨ ¸ V 1 OL Lij © miis ¹ © vis ¹ © dii ¹ § ps · (8) V 1 OCCij V ln ¨ sj ¸ V 1 Ei K US _ CHLij ¨p ¸ i © ¹ V 1 Ei K CHL _USij Hij Where Hij
V 1 eij eii .
Hence, the ratio of bilateral trade
§ mijs · ¸ is flows over intra-national trade flows in the industry s ¨ ¨m ¸ © iis ¹ explained by a list of observable variables: the relative value added
174
Maria Bas and Ivan Ledezma
§ v js · §d · ¸ , their relative distances ¨ ij ¸ , of partners in the industry ¨ ¨v ¸ ¨d ¸ © is ¹ © ii ¹ § p js · the relative prices ¨ s ¸ and the dummy variables Lij and C ij that ¨ pi ¸ © ¹ indicate if the pair (ij) shares the same language and has a common border, respectively. The part of missing trade not explained by these determinants is attributed to the fact that the exchanges take place between two particular and different nations. In that sense, the border effects coefficients reveal all market access difficulties for each specific trade flow. Other than "standard" trade policies, these difficulties might reflect consumers' preferences for domestic goods relative to foreign ones “home market bias”. As long as the latter might be interpreted, at least in part, as the outcome of history and political efforts, we consider them as a part of the measure of trade integration. In the regressions, we drop the constant and incorporate both dummy variables, each for one direction of trade flows. Thus, their coefficients can be directly interpreted as the border effect of each combination. For example, the exponential of the coefficient of US _ CHLij multiplied by -1, exp V 1 E i K , indicates the difficulty for Chilean exporters in accessing US markets. As in Fontagné et al. (2005) we use aggregate prices instead of industry-level prices to address the possible endogeneity problem that might arise in the estimation of equation (8). This issue is associated to the simultaneity in the determination of prices and output in a monopolistic competition framework. Prices at national level should be less likely correlated with profit maximization at the industry level. We also consider relative wages at the industry level. This alternative measure of relative prices is used to control for potential asymmetries in technology. More productive industries are expected to have lower border effects. We also run regressions using lagged values of relative value added and relative prices to reduce the risk of correlation between both sides of equation (8) in the nominal valuation of flows. When estimating equation (8) we should consider a representative set of countries trading with Chile. Looking at the aggregated trade flow data of the Economic Commission for Latin America and the Caribbean (ECLAC) between 1990 and 1999, the main
Trade Integration in Manufacturing: The Chilean Experience
175
destination of Chilean manufacturing exports are Latin America (LA), the United States (US) and the European Union (EU) (see Figure 2). During the same period, most manufacturing imports of Chile came also from these countries. Figure 2: Chile´s major trading partners Export of manufactured goods by country of destination (in per cent of total exports) 1990
13% 33%
15% 18% Latin America
21%
US EU 1999 9% 9%
11%
Asia
Others
Asia
Others
52%
19% Latin America
US
EU
Imports of manufactured goods by country of origin (in per cent of total imports) 1990 9%
19%
17% 23% 32% Latin America US 9%
EU 1999 9%
11% 19% Latin America US
Asia
Others
52%
EU
Asia
Others
176
Maria Bas and Ivan Ledezma
B. Data We use the Trade and Production Database constructed by CEPII. Its main source is the Trade and Production Database constructed by Nicita and Olarreaga (2001) from the World Bank. Since this database shows many missing values for production variables in recent years, CEPII extended it using production variables of UNIDO and OECD STAN for OECD members. CEPII also completed trade data with their international trade database BACI. The final dataset provides information on value added, export and import trade flows, origin and destination countries, wages and labour at the 3-digit industry level (ISIC Rev-2) over the period 1976-1999 for 67 developing and developed countries. Price indices are from Penn World Table as price level of GDP expressed relative to United States. Detailed intra-national trade flows for our sample of countries are not available. Following the work of Wei (1996), intra-national trade is computed as output minus exports. This requires an appropriate measure of internal distance that should take into account economic activity to weight internal regions (Head and Mayer, 2001). For distance variables, contiguity and common language, we use the CEPII database of internal and external distances. CEPII uses specific city level data in order to compute a matrix of distance including the geographic population density for each country. Similarly, distance between two countries are measured based on bilateral distance between cities weighted by the share of the city in the overall country's population. Finally we can include data for Chile, the US, nine members of the European Union throughout the whole period 1979-1999 (Germany, France, Great Britain, Italy, Belgium, Luxembourg, Ireland, Netherlands and Denmark) and for six Latin American partners (Bolivia, Argentina, Brazil, Mexico, Uruguay and Venezuela) III.
Estimation results
Equation (8) is first estimated using country-level data (all manufacturing) and then splitting the sample by industries. Thus firstly, we analyze the aggregated evolution of Chilean trade integration. We run pooled regressions and estimate border effects over all industries and years and also within different periods. Secondly, we take 2-digit industry-level data to explain differences in the degree of trade openness across sectors. From this estimation we obtain the
Trade Integration in Manufacturing: The Chilean Experience
177
global border effect measure for each industry as a weighted average of all trading partners.
A. Border effects at the country level Table 1 presents the results with country level data, i.e. across all industries and years. All regressions are run using OLS and Hubert and White standard errors to control for the correlation in the error term in equation (8). Although this theoretically founded equation suggests an elasticity of the relative value added equal to 1, we do not constrain this parameter in the empirical implementation. Column [1] reports regressions using current values of independent variables and aggregated prices. As predicted by theory, the coefficient of relative value added is significant, positive and close to one. The coefficient of relative prices is significantly negative. However, similarly to Fontagné et al. (2005) its value is smaller than the expected elasticity of substitution. The effect of contiguity is also significant and positive, indicating that having a common border facilitates trade. Surprisingly, the coefficient of the dummy variable indicating whether the partners share a common language is negative. This can be explained by the countries present in the sample. Relative to Latin American countries, which share the Spanish language, EU and US international flows are more important, offsetting the effect of language. In column [2] we use relative wages at the industry level instead of relative price at national level as explanatory. In the monopolistic competition framework relative prices equal relative wages. More importantly wages take into account labour productivity differentials among partners. Their effect turns out to be similar to relative prices but with a lower coefficient, which might reflect labour market asymmetries. Column [3] in Table 1 introduces the lag of relative value added and relative prices to address potential correlation between both sides of the equation stemming from the valuation of nominal trade flows and output. Results remain almost unchanged. A Similar conclusion can be obtained from column [4], which considers lagged relative wages. The coefficients of border effects estimates indicate that, on average, during the period 1979-1999 Chilean exporters faced higher barriers to enter the European Union than the barriers faced by European's exporters to sell in Chilean markets (5.386 versus 4.30). A similar asymmetry is found in flows regarding the US (4.061 against 3.897). These first results indicate that, on average, Chilean
178
Maria Bas and Ivan Ledezma
exporters experienced more difficulties to access foreign markets in the period than their partners faced in the Chilean market. The larger border effect found in the case of Latin American countries illustrate that distance, language and contiguity are not the only source of barriers. Although we find higher coefficients, these estimates are in line with Fontagné et al. (2005) for the US and the EU. Inside the EU trade is reduced by a factor of 14.87 (exp(2.73), column [1]) when crossing the border of European nations. Table 1: Border effects at the country level (to be continued)
Rel. VA
[1]
[2]
0.822***
0.829***
(0.007) Rel Price
[3]
[4]
(0.007)
-1.012*** (0.066)
Rel Wage
-0.742*** (0.013)
Rel.VA (t-1)
0.815*** (0.008)
Rel Price (t-1)
0.819*** (0.007)
-1.122*** (0.068)
Rel Wage (t-1)
-0.743*** (0.012)
Rel Distance contiguity Language EU_CHL CHL_EU US_CHL
-0.827***
-0.776***
-0.819***
-0.754***
(0.025)
(0.025)
(0.026)
(0.026)
0.876***
0.990***
0.890***
1.035***
(0.035)
(0.036)
(0.037)
(0.038)
-0.350***
-0.360***
-0.355***
-0.355***
(0.033)
(0.029)
(0.035)
(0.030)
-5.386***
-5.783***
-5.190***
-5.582***
(0.138)
(0.131)
(0.146)
(0.138)
-4.239***
-4.269***
-4.079***
-4.224***
(0.124)
(0.112)
(0.128)
(0.115)
-4.061***
-4.399***
-4.071***
-4.381***
(0.114)
(0.105)
(0.116)
(0.106)
Trade Integration in Manufacturing: The Chilean Experience
179
Table 1 (continued): Border effects at the country level [1] -3.897***
[2] -3.822***
[3] -3.828***
[4] -3.860***
(0.113)
(0.105)
(0.118)
(0.110)
EU_EU
-2.723***
-2.836***
-2.744***
-2.893***
(0.054)
(0.054)
(0.057)
(0.057)
LA_CHL
-5.120***
-4.946***
-4.944***
-4.828***
(0.092)
(0.093)
(0.096)
(0.097)
-4.765***
-5.200***
-4.622***
-5.099***
(0.103)
(0.099)
(0.107)
(0.102)
-5.347***
-5.663***
-5.276***
-5.674***
(0.079)
(0.077)
(0.083)
(0.081)
-5.020***
-5.609***
-5.141***
-5.726***
(0.120)
(0.110)
(0.126)
(0.115)
-4.774***
-4.913***
-4.679***
-4.945***
(0.098)
(0.088)
(0.103)
(0.092)
-4.508***
-4.403***
-4.359***
-4.406***
(0.091)
(0.081)
(0.095)
(0.084)
-3.752***
-4.738***
-3.856***
-4.811***
(0.095)
(0.085)
(0.099)
(0.087)
EU_US
-2.878***
-2.744***
-2.873***
-2.775***
(0.092)
(0.093)
(0.096)
(0.097)
US_EU
-2.685***
-2.883***
-2.703***
-2.938***
(0.070)
(0.068)
(0.072)
(0.071)
No. obs.
53278
63073
47843
57861
Adj. R2
0.912
0.911
0.912
0.911
CHL_US
CHL_AL LA_LA EU_LA LA_EU LA_US US_LA
Note: Huber White Standard errors in parentheses. *p < 0.10, **p < 0.05, ***p < 0.01
180
Maria Bas and Ivan Ledezma
Table 2: Evolution of border effects at the country level (to be continued) [79-82] Rel. VA
Rel Wage
Rel. Distance
Contiguity
Language
EU_CHL
CHL_EU
US_CHL
CHL_US
EU_EU
LA_CHL
CHL_LA
LA_LA
EU_LA
LA_EU
0.793***
[83-86] 0.835***
[87-90] 0.822***
[91-94] 0.834***
[95-99] 0.832***
(0.018)
(0.016)
(0.015)
(0.017)
(0.015)
-0.905***
-0.855***
-0.471***
-0.435***
-0.301***
(0.073)
(0.017)
(0.071)
(0.059)
(0.044)
-0.651***
-0.691***
-0.701***
-0.868***
-0.921***
(0.070)
(0.054)
(0.050)
(0.052)
(0.048)
1.014***
0.995***
1.068***
0.983***
0.779***
(0.084)
(0.081)
(0.078)
(0.074)
(0.070)
0.028
-0.061
-0.144**
-0.234***
-0.315***
(0.064)
(0.066)
(0.065)
(0.067)
(0.051)
-6.617***
-6.663***
-5.871***
-4.497***
-4.507***
(0.366)
(0.297)
(0.281)
(0.280)
(0.264)
-4.844***
-5.021***
-5.048***
-4.113***
-3.754***
(0.323)
(0.246)
(0.254)
(0.251)
(0.208)
-6.491***
-5.604***
-4.009***
-3.248***
-2.787***
(0.379)
(0.237)
(0.203)
(0.209)
(0.180)
-4.185***
-4.433***
-4.815***
-3.627***
-3.372***
(0.296)
(0.220)
(0.259)
(0.233)
(0.187)
-3.273***
-3.057***
-2.996***
-2.483***
-2.378***
(0.140)
(0.116)
(0.111)
(0.117)
(0.114)
-6.096***
-6.599***
-6.141***
-4.583***
-3.879***
(0.359)
(0.239)
(0.192)
(0.176)
(0.152)
-6.391***
-6.517***
-5.904***
-4.913***
-3.862***
(0.278)
(0.214)
(0.204)
(0.210)
(0.172)
-7.035***
-6.813***
-6.338***
-5.486***
-4.876***
(0.280)
(0.169)
(0.159)
(0.158)
(0.139)
-6.561***
-6.515***
-5.569***
-4.637***
-3.982***
(0.310)
(0.242)
(0.241)
(0.242)
(0.229)
-5.529***
-5.649***
-5.800***
-4.764***
-4.545***
(0.279)
(0.189)
(0.213)
(0.194)
(0.169)
Trade Integration in Manufacturing: The Chilean Experience
181
Table 2 (cont.): Evolution of border effects at the country level [79-82] LA_US
[83-86]
[87-90]
[91-94]
[95-99]
-4.635***
-5.212***
-5.441*** -4.120***
-4.115***
(0.250)
(0.177)
(0.202)
(0.158)
-6.292***
-5.440***
-4.354*** -3.827***
-3.076***
(0.218)
(0.182)
(0.195)
(0.184)
-3.207***
-3.175***
-3.189*** -2.339***
-2.253***
(0.260)
(0.201)
(0.188)
(0.186)
-3.886***
-3.307***
-3.002*** -2.550***
-2.123***
(0.170)
(0.147)
(0.140)
(0.145)
(0.136)
No. obs.
9110
12769
13950
11642
15602
Adj. R2
0.915
0.915
0.916
0.915
0.912
US_LA
EU_US
US_EU
(0.178)
(0.185)
(0.201)
Note: Huber White Standard errors in parentheses. *p < 0.10, **p < 0.05, ***p < 0.01
One possible explanation of these high average estimates may be the high protection of the Chilean economy in the early 1970s. To investigate this issue, we split the full sample in five periods. This time, regressions consider wage differentials, the specification in column [2] of Table 1. Clearly trade integration in recent years is higher than their initial level and the average of the period. On the other hand, our first results are confirmed: even though trade has become easier over time, Chile appears as more open to imports from Europe (CHL_UE) than Europe to Chilean exporters (UE_CHL). A similar asymmetry is found for trade flows with the US until the first half of 1980s. However, in the last two periods (19871999) the US has become more open to Chile than Chile vis-à-vis the US. Regarding Latin American countries trade barriers have evolved almost symmetrically, although the border effects still remain important. In the last period, trade between Chile and a Latin American country of the sample was reduced on average by a factor of 49 per cent (=exp(3.9)) in comparison with intra-national trade. Figure 3 summarizes the reduction of border effect over time by comparing the last period to the first one. The most important change in trade integration is found in the US market access for
182
Maria Bas and Ivan Ledezma
Chilean exporters (USÅCHL). Trade barriers faced by Chilean producers when reaching American markets have been reduced by around 57 per cent. In addition, considering differences instead of levels highlights asymmetries: Chilean border effects vis-à-vis the US have been reduced only by 19 per cent. Figure 3: Trade barriers reduction: 1979-82 versus 1995-99, in per cent 60% 50% 40% 30% 20% 10% -US
-L A
- US
-E U U S <-
EU<-
U S <-
-E U
-L A
-LA
L A<-
LA<-
EU<-
-EU
L A<-
< --L A
EU< -
C HL
S
- CH L
< --U
L A<-
C HL
U
-C H L
< -- E
U S <-
- CH EU< -
C HL
L
0%
B. Border effects at the industry level Finally, we estimate border effects at the 2-digit industry level to analyze differences across industries. To take into account the evolution over the full time period we run the regressions using a rolling window of four years. For instance, the border effect of the year 1982 is the estimate of the regression between 1979 and 1982. Figure 4 shows these estimates. All of them are significant, at least at the 5 per cent level. The dashed line depicts import border effects and the solid one those corresponding to exports. The figure summarizes the weighted average of border effect across partners using their share over total exports or imports.
Trade Integration in Manufacturing: The Chilean Experience
183
Figure 4: Development of border effects (in per cent) at the industry level Food (31)
Textil (32)
Exports
Imports
Exports
Wood (33)
1997
1994
1991
1988
1985
10.0 5.0 0.0 1982
1997
1994
1991
1988
1985
1982
10.0 5.0 0.0
Imports
Paper (34)
Exports
Imports
Exports
Chemicals (35)
1997
1994
1991
1988
1985
1997
1994
1991
1988
1985
1982
10.0 5.0 0.0
1982
10.0 5.0 0.0
Imports
Non Metallic Goods (36)
Exports
Imports
1998
1996
1994
1992
1990
1988
1986
1984
1998
1996
1994
1992
1990
1988
1986
1984
1982
Exports
1982
10.0 5.0 0.0
10.0 5.0 0.0
Imports
Machinery (38)
Basic Metals (37)
10.0
10.0 5.0 0.0
Exports
1998
1996
1994
1992
1990
1988
1986
1984
1998
1996
1994
Imports
0.0 1982
Exports
1992
1990
1988
1986
1984
1982
5.0
Imports
Difficulties of Chilean exporters accessing foreign markets were relatively similar across industries at the beginning of the 1980s. Reflecting the beginning of the active trade agreements agenda and political integration, most industries watched a downward trend in border effects which is especially pronounced during the 1990s. This is the case of wood, textile, plastic and machinery. Two important exporting industries, basic metals and food, experienced no reduction of export border effects. The first is, however, the most
184
Maria Bas and Ivan Ledezma
traditional exporting industry and its trade barriers were already low at the beginning of the period. On the other hand, the rather flat evolution of barriers of food exports relative to its consumption within the national border might reflect quality controls imposed by the EU and the US. Home biases are also likely associated with this type of industry. Once again, one observes the extent to which direct measures of trade such as tariffs do not reflect the whole picture of trade policies: export difficulties have considerably diminished in all industries during nineties even if tariffs were already very low. Figure 4 also shows the evolution of the weighted measure of industry level barriers faced by UE, LA and US to access Chilean markets (dashed line). In many industries market access difficulties increased during the first half of the 1980s (food, textile, wood, non metallic goods and machinery). This seems very consistent with the rise in import tariffs during this period and other discretionary measures to control the current account deficit during the crisis. As we consider a moving average, this tendency is observed even in the late 1980s, reflecting a lagged effect of protection. During the 1990s, import border effects have been reduced in almost all industries with the exception of basic metals. Tariffs attained a very low level at the beginning of the 1990s (11 per cent). This is also the decade in which the new strategy of trade integration based on several bilateral and multilateral trade agreements started. The reduction and convergence of import border effects observed for some industries (machinery, non metallic products, plastic and textiles) seems quite consistent with the agenda of Chilean trade policies. IV.
Conclusions
Using bilateral trade flow data of Chile with its principal trading partners for the period 1979-1999, we estimated the barriers to trade. Our econometric specification is theoretically founded on a monopolistic competition framework with increasing returns to scale, searching to explain inter-national import flows relative to intra-national ones. The estimates of international trade difficulties are thus relative to those faced by domestic producers reaching domestic consumers. Results are consistent with the agenda of trade integration followed by Chile. Namely, during the 1990s the estimated reduction of export barriers considerably slowed down. This implies
Trade Integration in Manufacturing: The Chilean Experience
185
asymmetries between the rapid reduction of export barriers and the tendency depicted by import's unilateral liberalization, especially regarding the US. At the industry level a certain convergence is observed for the last periods, reflecting the effects of free trades agreements. Overall, these results highlight a heterogeneity in trade barriers measures across industries, among export and import policies and over the time that is not captured using direct measure of trade policies. This heterogeneity could help to understand the effects of trade on manufacturing performances as it brings more variance concerning the outcome of trade policies. Further research should be oriented towards this issue.
References
Anderson, James, Van Wincoop, Eric (2004), Trade Costs, Journal of Economic Literature 42 (1), 691-751. Anderson, James, Van Wincoop, Eric (2003), Gravity with Gravitas: a Solution to the Border Puzzle, The American Economic Review 93 (3), 170-192. Baier, Scott, Bergstrand, Jeffrey (2001), The Growth of World Trade: Tariffs, Transport Costs and Income Similarity, Journal of International Economics 53 (1), 1-27. Bergstrand, Jeffrey (1989), The Generalized Gravity Equation, Monopolistic Competition, and the Factor-Proportions Theory in International Trade, The Review of Economics and Statistics 71 (1), 143-153. Cabezas, Mabel (2003), Tratado de libre comercio entre Chile y Estado unidos: revision de estudios que cuantifican su impacto, Central Bank of Chile Working Papers No. 239. Chumacero, Romulo, Fuentes, Rodrigo, Schmidt–Hebbel, Klaus, (2004), Chile's free trade arrangements: How big is the deal?, Central Bank of Chile Working Papers No. 264. Dixit, Avinash, Stiglitz, Joseph (1977), Monopolistic Competition and Optimum Product Diversity, The American Economic Review 67 (3), 297-308.
186
Maria Bas and Ivan Ledezma
Feenstra, Robert (2003), Advanced International Trade: Theory and Evidence, Princeton University Press. Fontagné, Lionel, Mayer, Thierry, Zignago, Soledad (2005), Trade in the triad: how easy is the access to large markets?, Canadian Journal of Economics 38 (4), 1402-1430. Harrigan, James (1996), Openness to trade in manufactures in the OECD, Journal of International Economics 40 (1-2), 23-39. Harrison, Glenn, Rutherford, Thomas, Tarr, David (2003), Chile's regional arrangements: the importance of market access and lowering the tariff to six per cent, Central Bank of Chile, Working Papers No. 238. Head, Keith, Mayer, Thierry (2001), Effet frontière, intégration économique et "Forteresse Europe", Document de travail CEPII No 2001-06, Paris. Herzer, Dierk, Nowak-Lehmann, Felicitas, Vollmer, Sebastian (2007), The Free Trade Agreement Between Chile and the EU: Its Potential Impact on Chile's Export Industry, Applied Econometrics and International Development 7 (1), 97-120. Krugman, Paul (1980), Scale Economies, Product Differentiation, and the Pattern of Trade, The American Economic Review 70 (5), 950-59. McCallum, John (1995), National Borders Matter, The American Economic Review 85 (3), 615-23. Nicita, Alessandro, Olarreaga, Marcelo (2001), Trade and production 1976-99, World Bank Policy Research Working Paper Series 2701, The World Bank. Redding, Stephen, Venables, Anthony (2004), Economic Geography and International Inequality, Journal of International Economics 62 (1), pp. 53-82. Wei Shang-Jin (1996), Intra-National Versus International Trade: How stubborn are nations in Global Integration?, NBER Working Papers No. 5531.
Inmaculada Martínez-Zarzoso, Felicitas Nowak-Lehmann D. and Sebastian Vollmer1
Comparing French and Spanish Exports to Emerging and Developed Markets Abstract In this paper we assess the current relevance of Ricardian theory. Relative prices, labor costs, and productivity are evaluated as determinants of a country’s international competitiveness at the industry level. Working with detailed data on unit values and with industry data on productivity, we empirically implement a MacDougall-type model for Spanish and French trade to Brazil, China, Japan, and the U.S.. The period under study is 1980 to 2001 and we distinguish in our analysis between homogenous, referencepriced, and differentiated goods. Our results indicate that Ricardian theory is currently only valid for explaining trade with developing countries while other factors are of importance for developed economies. Overall price competitiveness is of importance, but for differentiated goods, factors distinct from prices seem to determine export success. I. Introduction As a consequence of the globalization process (progressive deregulation of trade achieved in multilateral trade rounds, regional and bilateral trade agreements, and integrated production systems), trade flows have multiplied by 22 times since 1970, much more than the GDP during the same period. International trade flows have changed as dramatically in content and direction over the past three decades as they have in size. What is the response of national economies to globalization in 1
The authors acknowledge financial support from the Spanish Ministry of Education (SEJ 2007-67548). Sebastian Vollmer acknowledges financial support from the Georg Lichtenberg program “Applied Statistics and Empirical Methods”. Inmaculada Martínez-Zarzoso acknowledges financial support from Fundación Caja CastellónBancaja and Generalitat Valenciana (P1-1B2005-33, ACOMP07/102).
188
Inmaculada Martínez-Zarzoso, Felicitas Nowak-Lehmann D., Sebastian Vollmer
terms of trade? International trade theories (classical or new) predict that increasing globalization is associated with a higher production concentration of certain economic activities, and therefore increasing specialization, according to comparative advantage and economies of scale criteria. Contrary to the theoretical predictions (Krugman, 1991), recent empirical evidence shows that trade in the extensive margin (a wider set of goods) is the dominant trend for large economies (e.g., Hummels and Klenow, 2005). Although much attention has been given to the role of productivity differences in influencing international competitiveness, this relationship has scarcely been investigated within the framework of the classical Ricardian model. Since the early studies by MacDougall (1951, 1952), Stern (1962), and Balassa (1963), only a few authors have recently evaluated the empirical validity of the Ricardian model. Golub and Hsieh (2000) assessed the contemporary relevance of the classical model for US trade over the period 1970-1992. They found some evidence supporting the theory, but much of the sectoral variation in trade remained unexplained. Choudhri and Schembri (2002) used a modern adaptation of the Ricardian model, which incorporates monopolistic competition and derived a MacDougall-type relationship. They tested this relationship for Canada and the U.S. using panel data for 1966 through 1990 for forty industries. Their results also support the validity of the Ricardian model, although other factors, such as trade liberalization, also play an important role in explaining market shares. This paper extends and updates the existent literature using a different set of countries and years. Since most of the empirical evidence in this field is related to U.S. international trade, we aim to extend the evidence to other countries, and specifically to North-South trade. In the remainder of the paper, Section II briefly presents the Ricardian model and the empirical implementation of the model. Section III presents some results before Section IV concludes. II. Modelling relative export strength A. The Ricardian Model According to the Ricardian model of free trade, countries tend to export those goods which have the lowest relative costs in autarky. In its simplest form, comparative advantage is defined in terms of unit labor requirements in a world of two goods and two countries.
Comparing French and Spanish Exports to Emerging and Developed Markets
189
Assuming that labor is the only factor of production, the supply of labor is fixed in each country and perfect competition prevails in all markets; Country 1 has a comparative advantage in producing Good i, compared to Country 2 and Good j, if it can produce Good i with less labor relative to Good j, compared to Country 2. Thus,
ai1 ai2 a 1j a 2j
(1)
It can be shown that world output increases if one or both countries specialize in producing the good in which they have comparative advantage. The Ricardian model can easily be generalized to multiple goods, i=1,…,N. A ranking can be constructed over the N goods’ relative labor requirements in the two countries. The new formulation in terms of Country 1’s labor requirements is given by,
a 1N a11 a 12 ... 2 a12 a 22 aN
(2)
According to the theory, Country 1 specializes in goods that lie to the left in this chain, whereas Country 2 specializes in goods that lie to the right. Assuming free trade, a unique break point exists that determines the patterns of specialization in both countries. Although this result does not fully specify the pattern of trade, as in the two-goods case (since a number of goods may not be exported by one or both countries), it nevertheless shows some role for comparative advantage. If Countries 1 and 2 are any two countries in the world, in a world of many countries, it can still be stated that with free trade all of the goods exported by Country 1 and not exported by Country 2 will lie on the left of goods exported by Country 2 in this chain. Therefore, Equation 2 could be used to make partial statements about patterns of trade. In a world of many goods and countries, specialization is associated somehow with low unit labor requirements. B. Data sources and empirical implementation. The main problem faced by researchers when they attempt to test the Ricardian theory is that autarky prices are not observable; hence the theory of comparative advantage cannot be directly tested. However, there are other factors that may be observable and that help to explain which goods countries trade; autarky prices then could be explained with country-specific characteristics.
190
Inmaculada Martínez-Zarzoso, Felicitas Nowak-Lehmann D., Sebastian Vollmer
The main data sources are the Groningen Center for productivity and labor costs at the industry level and the UN COMTRADE database for disaggregated exports in value and volume. To test the Ricardian model, we perform a panel analysis of Spanish exports relative to French exports in the period 1980 to 2001. Table 1 summarizes some descriptive statistics on labor productivity and labor costs for France relative to Spain at the industry level. Whereas a figure greater than one indicates an advantage in productivity for France relative to Spain, figures greater than one in labor costs and unit labor costs indicate a disadvantage for France relative to Spain. For all sectors we observe relative advantages in productivity (value added per hour) and relative disadvantages in labor costs for France. The unit labor costs summarize both indicators in one figure. Here we find a mixed picture, relative advantages for France in most of the more differentiated sectors and relative advantages for Spain in the other sectors. On average, Spain has a slight advantage over France in total manufacturing measured in unit labor costs. Following MacDougall (1951, 1952), Stern (1962), and Balassa (1963), export ratios are used as a measure of trade. As in Balassa, we use exports to third markets. The independent variables considered are: relative productivity, relative labor costs, and relative unit values. We set up the following specifications to capture the model described above: § X ijkt · § UV · ¸ D j E j ln¨ ijkt ¸ F j ln§¨ aikt ·¸ G j ln§¨ wikt ·¸ H jkt (3) ln¨ ¨w ¸ ¨a ¸ ¨ X ljkt ¸ ¨ UVljkt ¸ © lkt ¹ © lkt ¹ © ¹ © ¹
§ X ijkt · § UV · ¸ D j E j ln¨ ijkt ¸ O j ln§¨ ulcikt ·¸ P jkt , (4) ln¨ ¨ ulc ¸ ¨ X ljkt ¸ ¨ UVljkt ¸ © lkt ¹ © ¹ © ¹ where Xijkt (Xljkt) denotes exports from Country i (l) to Country j, for Sector k in Period t. UVijkt and UVljkt denote French (i) and Spanish (l) export unit values to Destination j, for Sector k in Period t. aikt and aljkt denote French (i) and Spanish (l) labor productivity for sector k in period t. wikt and wlkt denote French (i) and Spanish (l) labor compensation per employee for Sector k in Period t. ulc denotes unit labor costs and is calculated from a and w. As destination markets j, we chose Japan, the U.S., Brazil, and China, two large developed economies, as well as two important emerging markets.
Food, drink & tobacco Textiles Clothing Leather and footwear Wood & products of wood and cork Pulp, paper & paper products Printing & publishing Chemicals Rubber & plastics Non-metallic mineral products Basic metals Fabricated metal products Mechanical engineering Office machinery Insulated wire Other electrical machinery and app. nec. Electronic valves and tubes
Relative value added per Relative labor costs per Relative unit labor costs hour hour Min. Avg. Max. Min. Avg. Max. Min. Avg. Max. 1.267 1.434 2.072 1.344 1.591 1.790 0.838 1.122 1.351 1.265 1.466 1.665 1.457 1.751 1.922 1.025 1.199 1.383 1.091 1.409 1.774 1.659 1.966 2.288 1.147 1.413 1.723 1.350 1.770 2.685 1.720 1.984 2.260 0.740 1.162 1.428 1.493 1.671 2.461 1.430 1.715 2.082 0.804 1.033 1.324 0.809 1.129 1.463 1.296 1.574 1.738 0.906 1.422 2.018 1.109 1.379 1.981 1.403 1.725 2.333 0.937 1.276 1.554 1.299 1.574 1.968 1.273 1.625 1.797 0.863 1.040 1.335 0.916 1.087 1.328 1.050 1.408 1.644 1.016 1.305 1.656 1.075 1.176 1.296 1.391 1.777 1.991 1.123 1.518 1.787 1.050 1.354 1.638 1.202 1.426 1.590 0.734 1.079 1.512 1.759 2.124 2.663 1.415 1.769 2.070 0.718 0.839 0.980 1.285 1.506 1.748 1.313 1.578 1.769 0.903 1.055 1.377 1.931 3.192 9.439 1.516 2.239 3.055 0.292 0.795 1.293 1.195 1.586 2.066 1.184 1.483 1.820 0.720 0.947 1.270 1.408 1.662 1.862 1.321 1.606 1.861 0.834 0.968 1.113 1.684 2.348 3.088 1.267 1.671 2.021 0.516 0.741 1.155
Table 1: Productivity and labor costs of France relative to Spain (to be continued)
Telecommunication equipment Radio and television receivers Scientific instruments Other instruments Motor vehicles Building and repairing of ships and boats Aircraft and spacecraft Railroad equipment and transport equip. nec. Furniture, miscellaneous manufacturing; recycling Total manufacturing 1.474 1.741
1.495 1.404
1.788 1.707
2.080 1.893
1.133 0.996
0.468
1.314 1.132
0.854
1.473 1.299
1.747
1.363 1.512
2.331
1.176 1.400
1.508
1.073
1.909
0.978
2.783
Relative labor costs Relative unit labor costs per hour Min. Avg. Max. Min. Avg. Max. 0.931 1.355 1.809 0.464 0.664 0.867 0.824 1.449 2.486 0.427 0.801 1.439 1.512 1.859 2.172 0.625 0.864 1.137 0.901 1.323 2.465 0.626 0.893 1.307 1.047 1.364 1.655 0.920 1.396 1.970 0.730 1.882 2.994 0.416 1.122 1.618 1.181 1.895 2.617 0.434 0.923 1.502
Relative value added per hour Min. Avg. Max. 1.130 2.116 3.165 1.455 1.842 2.515 1.809 2.198 2.882 1.127 1.474 1.886 0.691 1.006 1.457 0.996 1.795 3.535 1.077 2.423 4.910
Table 1 (cont.): Productivity and labor costs of France relative to Spain
Comparing French and Spanish Exports to Emerging and Developed Markets
193
One should note that the unit values differ across sectors and destination markets, whereas the productivity and labor-cost data are constant across destination markets and only differ across sectors. Following Rauch (1999), we classify sectors into three different groups, namely homogeneous (Rauch 1), reference-priced (Rauch 2), and differentiated goods (Rauch 3). In our estimations we include dummies for the different groups into Equations 3 and 4 or restrict our sample to one of the three groups to acknowledge the differences between these types of goods. We would expect that price competitiveness is less important for differentiated goods than it is for homogeneous or reference-priced goods. Other factors like quality, variety, or uniqueness should be more important for these types of goods. III.
Determinants of French and Spanish relative export strength
In the empirical application, we simplify the terms used in Equations 3 and 4 which have France in the numerator and Spain in the denominator. lxv stands for relative (France over Spain) export strength in logs, luv is utilized for relative unit values in logs, lw is relative labor compensation, and lva is relative productivity in logs. j characterizes the destination market, k stands for sector, and t stands for time. We obtain the following Equations 5 and 6: lxv jkt D j E j luv jkt F j lva jkt G j lw jkt H jkt (5) lxv jkt
D j E j luv jkt O j lulc jkt H jkt .
(6)
To control for cross-correlation between destination markets j, we estimate specifications (5) and (6) respectively as a system, with one equation for each destination market using seemingly unrelated regression (SUR). In prior estimations, we allow for countryspecific constants and country-specific coefficients. However, testing for equality of the coefficients using a Wald test indicates that the differences between the coefficients are not significant in all cases. We therefore estimate Equations 5 and 6 with common coefficients and country-specific constants. Autocorrelation is addressed by the inclusion of AR(1) terms. It appears that het-
194
Inmaculada Martínez-Zarzoso, Felicitas Nowak-Lehmann D., Sebastian Vollmer
eroskedasticity does not affect the estimated coefficients; all SUR results presented are robust compared to GLS approaches.2 Table 2 reports the results for French, relative to Spanish, exports to four destination markets: the U.S., Brazil, Japan, and China. The period under study is 1980 to 2001. At a first glance these results support the Ricardian theory of comparative advantage; the coefficients for unit labor costs, labor costs, and value added are significant and show the expected sign. The coefficient of the unit values is significant with the expected sign as well, but it is rather small. The negative and significant coefficients of the Rauch 1 and Rauch 2 dummies indicate that France has an advantage over Spain in exporting differentiated goods compared to homogenous and reference-priced goods. Compared to earlier studies, the explanatory power of our estimates is rather high, which might be due to the fact that this study, in contrast to previous ones, evaluates the time-series properties of the data. In a next step we estimated the model for reference-priced and differentiated goods separately. The results are shown in the second and third columns of Table 3. Since our productivity data mainly include manufacturing, there is only a relatively small number of observations available for homogeneous goods, excluding most agricultural products. Consequently, we prefer not to draw conclusions from a subsample, including only Rauch 1 goods. For reference-priced goods, we find that that labor costs and unit values have a greater impact than they have for all goods, whereas productivity turns out to be insignificant. In contrast to differentiated goods, unit values have a positive sign, indicating that other factors apart from price competitiveness seem to play a role. While labor costs are only weakly significant, both labor costs and productivity carry the expected signs.
2
Models 5 and 6 were also estimated using weighted least squares. Results are available upon request from the authors.
Comparing French and Spanish Exports to Emerging and Developed Markets
195
Table 2: Determinants of French Export Strength Relative to Spanish Export Strength for Brazil, China, Japan, and the U.S. Estimation method: seemingly unrelated regression Included observations: 9743 Total system (unbalanced) observations 21058 Dependent variable: lxv Coefficient Prob. Coefficient Prob. lulc -0.229 0.022 lw -0.231 0.117 lva 0.228 0.060 luv -0.039 0.000 -0.039 0.000 AR(1) 0.749 0.000 0.749 0.000 Rauch 1 -1.478 0.000 -1.479 0.000 Rauch 2 -0.429 0.000 -0.429 0.000 Constant Brazil 1.066 0.000 1.068 0.000 Constant China 1.482 0.000 1.484 0.000 Constant Japan 2.304 0.000 2.305 0.000 Constant U.S. 1.792 0.000 1.793 0.000 Brazil Observations: 4352 Adjusted R-squared 0.513 0.513 Durbin-Watson 2.166 2.166 China Observations: 2468 Adjusted R-squared 0.310 0.309 Durbin-Watson 2.394 2.394 Japan Observations: 6060 Adjusted R-squared 0.641 0.641 Durbin-Watson 2.068 2.068 U.S. Observations: 8178 Adjusted R-squared 0.647 0.647 Durbin-Watson 2.158 2.158 Note: All variables except dummies are in natural logs and in relative terms (France relative to Spain). lulc denotes unit labor cost, lw denotes labor compensation, lva denotes value added per hour, and luv denotes unit values. Autocorrelation was corrected by adding an AR(1) to the model specification.
196
Inmaculada Martínez-Zarzoso, Felicitas Nowak-Lehmann D., Sebastian Vollmer
Table 3: Determinants of French Export Strength Relative to Spanish Export Strength for Brazil China, Japan, and the U.S. (Rauch 2 and Rauch 3) Estimation method: seemingly unrelated regression Included observations: 2252 7146 Total system observations: 4739 15803 Dependent variable: lxv Rauch 2 Rauch 3 Coefficient Prob. Coefficient Prob. lw -1.139 0.006 -0.283 0.080 lva -0.329 0.285 0.412 0.002 luv -0.283 0.000 0.022 0.043 AR(1) 0.721 0.000 0.734 0.000 Constant Brazil 1.679 0.000 0.956 0.000 Constant China 1.394 0.000 1.521 0.000 Constant Japan 2.453 0.000 2.270 0.000 Constant U.S. 2.147 0.000 1.692 0.000 Brazil Observations: 1076 3198 Adjusted R-squared 0.540 0.491 Durbin-Watson 2.043 2.187 China Observations: 599 1832 Adjusted R-squared 0.230 0.325 Durbin-Watson 2.202 2.392 Japan Observations: 1263 4637 Adjusted R-squared 0.617 0.629 Durbin-Watson 2.100 2.025 U.S. Observations: 1801 6136 Adjusted R-squared 0.623 0.655 Durbin-Watson 2.055 2.121 Note: All variables except dummies are in natural logs and in relative terms (France relative to Spain). lulc denotes unit labor cost, lw denotes labor compensation, lva denotes value added per hour, and luv denotes unit values. Autocorrelation was corrected by adding an AR(1) to the model specification.
Comparing French and Spanish Exports to Emerging and Developed Markets
197
Table 4: Determinants of French Export Strength Relative to Spanish Export Strength for Brazil and China Estimation method: seemingly unrelated regression Included observations: 5557 Total System Observations: 6820 Dependent variable: lxv Coefficient Prob. Coefficient Prob. lulc -0.688 0.002 lw -0.803 0.017 lva 0.642 0.007 luv -0.054 0.005 -0.054 0.005 AR(1) 0.650 0.000 0.650 0.000 Rauch 1 -2.476 0.000 -2.485 0.000 Rauch 2 -0.377 0.012 -0.382 0.011 Constant Brazil 1.221 0.000 1.298 0.000 Constant China 1.603 0.000 1.681 0.000 Brazil observations: 4352 Adjusted R-squared 0.515 0.515 Durbin-Watson 1.954 1.954 China observations: 2468 Adjusted R-squared 0.339 0.339 Durbin-Watson 2.234 2.234 Note: All variables except dummies are in natural logs and in relative terms (France relative to Spain). lulc denotes unit labor cost, lw denotes labor compensation, lva denotes value added per hour, and luv denotes unit values. Autocorrelation was corrected by adding an AR(1) to the model specification.
When we compare the results obtained from the reference-priced and differentiated goods sub-samples (Table 5) we find that competitiveness in labor costs and unit values is of major importance for reference-priced goods, while productivity turns out to be insignificant. For differentiated goods, both labor costs and productivity are significant and show the expected signs. In contrast, the unitvalue coefficient has a positive and weakly significant sign, indicating that it is not price competitiveness that explains export success for this type of goods.
198
Inmaculada Martínez-Zarzoso, Felicitas Nowak-Lehmann D., Sebastian Vollmer
Table 5: Determinants of French Export Strength Relative to Spanish Export Strength for Brazil and China (Rauch 2 and Rauch 3) Estimation method: seemingly unrelated regression Included observations: 1425 4009 Total system observations: 1675 5030 Dependent variable: lxv Rauch 2 Rauch 3 Coefficient Prob. Coefficient Prob. Lw -1.973 0.012 -0.767 0.041 Lva 0.447 0.391 0.752 0.005 Luv -0.368 0.000 0.041 0.064 AR(1) 0.636 0.000 0.639 0.000 Constant Brazil 1.858 0.000 1.154 0.000 Constant China 1.556 0.000 1.696 0.000 Brazil observations: 1076 3198 Adjusted R-squared 0.537 0.492 Durbin-Watson 1.854 1.976 China observations: 599 1832 Adjusted R-squared 0.262 0.349 Durbin-Watson 2.054 2.229 Note: All variables except dummies are in natural logs and in relative terms (France relative to Spain). lulc denotes unit labor cost, lw denotes labor compensation, lva denotes value added per hour, and luv denotes unit values. Autocorrelation was corrected by adding an AR(1) to the model specification.
The results for the two developing countries differ from what we find when we restrict the sample to Japan and the U.S. (Table 6). Neither the coefficients of unit labor costs nor productivity or labor costs are significant. Only price competitiveness seems to play a role, but the estimated coefficient is rather small. Hence, the estimation results for the developed countries in our sample do not support the theory of comparative advantage and specialization in production.
Comparing French and Spanish Exports to Emerging and Developed Markets
199
Table 6: Determinants of French Export Strength Relative to Spanish Export Strength for Japan and the U.S. Estimation method: seemingly unrelated regression Included observations: 9456 Total system observations: 14238 Dependent variable: lxv Coefficient Prob. Coefficient Prob. Lulc -0.132 0.229 Lw -0.117 0.467 Lva 0.143 0.291 Luv -0.034 0.002 -0.034 0.002 AR(1) 0.784 0.000 0.784 0.000 Rauch 1 -1.324 0.000 -1.322 0.000 Rauch 2 -0.472 0.000 -0.471 0.000 Constant Japan 2.258 0.000 2.245 0.000 Constant U.S. 1.752 0.000 1.740 0.000 Japan observations: 6060 Adjusted R-squared 0.643 0.643 Durbin-Watson 2.153 2.154 U.S. observations: 8178 Adjusted R-squared 0.648 0.648 Durbin-Watson 2.249 2.249 Note: All variables except dummies are in natural logs and in relative terms (France relative to Spain). lulc denotes unit labor cost, lw denotes labor compensation, lva denotes value added per hour, and luv denotes unit values. Autocorrelation was corrected by adding an AR(1) to the model specification.
IV.
Conclusions
The aim of our study was to examine the validity of the Ricardian model for different types of destination markets. Brazil and China are representatives of emerging/developing markets; Japan and the U.S. represent highly industrialized countries. Theory would lead us to expect more inter-industry (Ricardo type) trade between France and Spain and developing countries and to observe more intra-industry trade between France and Spain and industrialized countries. Furthermore, according to the theory, inter-industry trade is driven by price competitiveness factors, whereas intra-industry trade is
200
Inmaculada Martínez-Zarzoso, Felicitas Nowak-Lehmann D., Sebastian Vollmer
driven primarily by factors related to taste differences, product variety, and product quality. In fact, our empirical analysis indicates that Spanish exports to developing countries (Brazil and China), relative to French exports to those countries, can well be explained by the Ricardian model, i.e., by labor compensation and labor productivity (unit labor costs). In contrast, Spanish exports to developed countries, relative to French exports, are not so much determined by unit labor costs but by product characteristics. This conclusion is supported by the high proportion of intra-industry trade among industrialized countries. Products are imported because consumers of developed countries desire variety and are willing to pay more for a product with certain characteristics. The empirical evidence shows that the simpler model with common coefficients for destination markets provides more robust results than the model with destination-market-specific coefficients. However, there are some interesting differences in the coefficients when different types of products are investigated. Relative exports of products in the categories homogenous goods (Rauch 1) and reference-priced goods (Rauch 2) depend on price advantages and are therefore governed by price competitiveness factors. In contrast, relative exports of differentiated products (Rauch 3) are positively related to unit values. For this type of good, a higher relative price seems to be an indicator of higher quality or superior product properties, explaining why relative exports rise with increasing prices. In summary, the results in this paper add some evidence demonstrating that Ricardian theory is valid to explain North-South trade. Although price competitiveness is almost always an issue, other factors aside from price differences are probably more relevant in determining export success for differentiated goods. Consequently, “new” trade theories, related to monopolistic competition and economies of scale, are certainly more appropriate to explain trade among developed countries.
Comparing French and Spanish Exports to Emerging and Developed Markets
201
References
Balassa, B. (1963), An empirical demonstration of classical comparative cost theory, The Review of Economics and Statistics 4, 231-238. Choudhri, E. U., Schembri, L. L. (2002), Productivity performance and international competitiveness: An old test reconsidered, Canadian Journal of Economics 35 (2), 341-362. Golub, S. S., Hsieh, C. T. (2000), Classical Ricardian theory of comparative advantage revisited, Review of International Economics 8 (2), 221-234. Hummels, D., Klenow, P. J. (2005), The variety and quality of a nation's exports, American Economic Review 95 (3), 704-723. Krugman, P. (1991), Increasing Returns and Economic Geography, Journal of Political Economy 99 (3), 483-499. MacDougall, G. D. A. (1951), British and American exports: A study suggested by the theory of comparative cost. Part I, The Economic Journal 61 (December), 697-724. MacDougall, G. D. A. (1952), British and American exports: A study suggested by the theory of comparative cost, Part II, The Economic Journal 62 (September), 487-521. Rauch, J. E. (1999), Networks versus Markets in International Trade, Journal of International Economics 48, 7-35. Stern, D. (1962), British and American productivity and comparative costs in international trade, Oxford Economic Papers 14, 275-296.
Part 4: The EU and Asian Economies
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
Assessing the Complementarity between the EU and ASEAN Economies Abstract Using a post-Vinerian approach, the paper argues that the economic success of an ASEAN-EU FTA depends on the level of production and trade complementarity existing between the two regions: the greater the complementarity, the higher the scope for trade expansion between the two economies, the lower the costs induced by resource reallocation and the more successful the FTA. Based on a thorough analysis of the productive and trade structures in both regions, as well as on calculations of indices of revealed comparative advantages and of intra-industry trade, the paper concludes that evidence of a multi-faceted trade complementarity between the two regions renders the FTA all the more desirable and may also contribute to reduce the trade and investment asymmetry currently existing between the EU and ASEAN. I. Introduction With the now well-documented rise of regional production networks in East Asia,1 as well as the recent proliferation of free trade agreements (FTAs) negotiated among Asian economies, the EU seems to have lost ground in the region and the EU – ASEAN economic partnership has been to some extent sidelined, with ASEAN being today only the EU’s fifth major trading partner. One way to revive the old partnership2 is to design an ASEAN-EU FTA,
1
2
For evidence on the rising de facto integration of ASEAN with East Asia see Ando and Kimura (2003), Ng and Yeats (2003) and Athukorala and Yamashita (2005). The EC-ASEAN Trade and Cooperation Agreement was signed in 1980. The importance of the EC-ASEAN partnership was reaffirmed in September 2001 with the European Commission’s Communication "Europe and Asia: A Strategic Framework for Enhanced Partnership"
206
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
an objective that has recently been entrusted to the EU-ASEAN Vision Group.3 This move is moreover in line with the EU’s strategy to complement multilateralism with bilateral and interregional FTAs and with its rising awareness of the importance of Asia as an economic partner. The chapter argues that the economic success of an ASEAN-EU FTA depends on the level of production and trade complementarity existing between the two regions. Complementarity must be understood in the sense of the dissimilarity of productive structures and specialization patterns between the two regions. Using a postVinerian approach, the greater the complementarity, the higher the scope for trade expansion between the two economies, the lower the costs induced by resource reallocation and the more successful the FTA. The aim of this chapter is therefore to assess the degree of complementarity existing between ASEAN and the EU so as to identify the risk industries, sensitive sectors and promising areas in the design of the FTA. Economic complementarity will involve analyzing the productive and trade structures in both regions with due reference to revealed trade comparative advantage and intra-industry trade indices (Section 2). As a preliminary step, a brief overview of the general characteristics of the two partners is provided, with a view to highlighting the differences and similarities between the two regions (Section 3). Section 4 concludes. II. An overview of ASEAN – EU trade and investment relations A. Differences despite some similarities In spite of comparable population sizes,4 the two regions differ sharply on a number of counts. They first differ in terms of economic size, with EU-25 accounting for more than 30 per cent of world GDP and ASEAN for less than 2 per cent. Another major distinguishing feature between the two regions is the intensity of
3
4
and again in July 2003 with the communication ‘A New Partnership with South East Asia’. The Vision Group was established in 2005 by the EU Commission and ASEAN governments to study ways of strengthening and deepening the economic relations between the two regions. Some 545 million inhabitants for ASEAN-10 against 485 million inhabitants for EU-27 (World Bank, World Development Indicators 2006).
Assessing the Complementarity between the EU and ASEAN Economies
207
economic integration (intra-regional integration): intra-ASEAN trade and investment is low when compared with EU figures.5 IntraASEAN trade has been below 25 per cent of total ASEAN trade for many years, despite renewed efforts such as the entry into force of the ASEAN Free Trade Agreement (AFTA) in January 2002. Similarly, although intra-ASEAN FDI flows recovered from the aftermath of the Asian crisis, they have been stagnating at a low 12 per cent of total FDI flows into the region, and they would be even more modest in the absence of Singapore, which accounts for twothirds of intra-ASEAN FDI flows. Table 1: Intra economic integration: EU and ASEAN compared 1995 2005 in per cent Trade Intra-ASEAN exports 21* 23* Intra-ASEAN imports 17* 21* Intra EU-25 trade 66 65 Investment Intra-ASEAN FDI 16 12 Intra-EU-25 FDI (stocks) 62 Sources: for ASEAN: ASEAN Secretariat Figures, Jakarta. For the EU: EUROSTAT, Brussels. (*) Figures for 1993 and 2003.
Several reasons account for the relatively low level of economic integration in ASEAN: a low institutional commitment, weak policies, and, more importantly perhaps, a de facto integration of ASEAN with East Asia. Several empirical studies (Borrus et al., 2000; Matthews and Cho, 2000; and McKendrick et al., 2000) have highlighted the importance of MNEs forming regional and global production networks in the region, especially in textile and garments, automobiles and electronics.6 These studies argue that
5
6
Taking each individual country with respect to ASEAN as a whole shows that intra-ASEAN trade has increased between 1993 and 2003 for countries such as Indonesia, Thailand and Myanmar, albeit on the export side only, whereas a moderate increase is registered for the Philippines. In the case of Singapore, and in spite of its role as a trade platform for the ASEAN region, no increase is registered on the export side, whereas only a very slight increase is visible on the import side. More than any other developing region until recently, ASEAN has based its economic development on FDI. With the Asian financial
208
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
such market-driven East Asian integration with intra-firm, interfirm transactions, parent-subsidiary and other suppliers is more significant and rapid than government-to-government ASEAN integration schemes. By contrast, the EU epitomizes a high level of institutional and policy commitment, resulting in a high level economic integration. First, the EU is a well-structured economic union and the grouping speaks with one voice at trade and investment related international fora. In particular, the European Commission (EC) has exclusive competence in formulating and ensuring the implementation of the Common Commercial Policy, covering trade in all goods, and most services. Secondly, as highlighted above, the share of intra-regional trade and investment is higher than in any other economic region in the world. Also, in contrast to what can be observed in East Asia, intra-EU-15 economic integration was dominated until recently by horizontal intra-industry trade which primarily involved end-use differentiated products. This specific intra-regional pattern of trade could be explained by the relatively higher degree of homogeneity prevailing in EU-15. However, as a result of the latest enlargement of Eastern European countries, this pattern is gradually changing and vertical intra-industry trade has been gaining ground since the 1990s. According to Ando and Kimura (2005), empirical evidence suggests that a rapid fragmentation of production has occurred between the UK, Germany and France on the one hand and some of the new Member States (Czech Republic, Hungary, Poland and Slovakia) on the other. As a result of the rising heterogeneity now prevailing in EU-25, one way in which the two regions share common characteristics is in terms of their relative intra-regional productive strengths (Andreosso-O’Callaghan et al., 2006). Both ASEAN and EU countries display extremely diverse production structures and productivity levels. In ASEAN, Singapore is relatively specialized in non-distributive trade services, and Brunei in primary activities. Of the remaining ASEAN countries, there is a clear division between the less developed ASEAN members (Myanmar, Laos, and Cambodia) and the others (Thailand, Malaysia, the Philippines and Indonesia), which are more engaged in manufacturing activities. Vietnam is catching up rapidly with these five countries. Indonesia crisis and the dot-com crash, FDI inflows as a share of GDP were in the early 2000s at roughly the same level as in the mid to late 1980s.
Assessing the Complementarity between the EU and ASEAN Economies
209
stands out as being more equipped to withstand increased competition in the electrical machinery sector (an important sector for ASEAN). Office & computing machinery could be a problem area in the case of the Philippines and Malaysia, since this is an industry that displays below national average productivity levels. On the EU side, the new Member States from Eastern Europe (such as Estonia, Poland and Slovenia) tend to be relatively specialized in traditional industries such as textile and clothing, basic metals and coke, although a few of them have managed to climb quickly the technology ladder (such as Hungary and the Czech Republic). For some EU countries, the specialization patterns resemble therefore those of some of the ASEAN countries. ‘Advanced’ EU countries are specialized in knowledge-intensive manufacturing industries (Finland and Sweden), and in knowledge intensive services (UK, France, Luxembourg). At a country level of analysis, Germany and the UK are the only two countries that are consistently specialized in high or medium tech industries. However, the overall distinguishing feature of the EU, when compared with ASEAN, is its relative specialization in knowledgebased services, such as financial services and business services. In terms of the risks and challenges brought about by further structural change, emanating for example from an ASEAN-EU FTA, one can note: (i) Motor Vehicles and machinery equipment are important industries for both regions in toto. (ii) The EU is weak in office and computing machinery and is insufficiently specialized in high tech activities (when compared with the US). (iii) The new EU Members States, as well as Portugal and Italy, still excessively rely on textile and clothing. B. A highly asymmetric partnership Trade: The relative intensity of trade between the two regions presents a striking asymmetry. Although the ASEAN share of total extra-EU exports reached 6.5 per cent in 1996 (Strange et al. 2000), recent data from the European Commission shows that extra-EU-25 exports of manufactured products (SITC 0-9) to the ASEAN-10 reached 4.15 per cent of the total (i.e. 48 422 million euro) in 2006. As a result, ASEAN ranks fifth, in terms of the EU-25 major export markets, after the US (23 per cent), Switzerland (7.4 per cent), Russia (6.2 per cent) and China, excluding Hong Kong (5.4 per cent), but now ahead of Turkey and Japan. The situation does not
210
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
differ much on the import side: with a share of 5.8 per cent, the ASEAN-10 was ranked as the fifth largest extra-EU-25 import source in 2006, after China (excluding Hong Kong, 14.2 per cent), the US (13.1 per cent), Russia (10.1 per cent), and Norway (5.9 per cent), but ahead of Japan (5.7 per cent). Table 2: EU merchandise trade with ASEAN (Million Euro) Share of EU imports Share of EU exports from ASEAN in to ASEAN in total Year Imports Exports Balance total EU imports EU exports (in %) (in %) 2001 2002 2003 2004 2005 2006
70,791 67,731 65,770 69,105 71,137 78,190
7.20 7.19 6.99 6.69 6.01 5.79
43,842 40,513 39,247 42,880 45,012 48,422
4.91 4.50 4.47 4.45 4.24 4.15
-26,949 -27,218 -26,523 -26,224 -26,125 -29,768
Source: Directorate General (DG) Trade, Brussels, August 2007.
In contrast, the EU is the second largest export destination for ASEAN countries, after the US. Based on European Commission DG Trade data, the EU represents 12.9 per cent of ASEAN total exports in 2006 (down from close to 15 per cent in 2002, however). On the import side, the EU is now the third largest import partner of ASEAN (with a share of 10.3 per cent in total ASEAN imports in 2006), after Japan and China, but ahead of the US.7 This aggregate picture hides contrasted patterns across countries. A closer look at ASEAN by country shows that trade with the EU is of minor significance in the case of some of the new ASEAN members (such as Myanmar) as well as for Brunei, while the EU is a more important trade partner for Singapore, Malaysia and Thailand for instance.8 The great bulk of trade between the EU and ASEAN in both directions is in manufactured goods, and within the manufacturing category, in machinery, mechanical appliances and electrical equipment (code 84-85). According to European Commission DG 7
8
Japan’s first place is explained by the density of its systems of production in the region. It is noteworthy that the EU has overtaken the US. Based on IMF Direction of Trade Statistics 2006.
Assessing the Complementarity between the EU and ASEAN Economies
211
Trade figures, the first five commodities that the EU exported to ASEAN in 2006 were machinery, mechanical appliances and electrical equipment (chapters 84-85, with 47.4 per cent), chemicals (chapters 28-38, with 11.6 per cent), motor vehicles (chapters 8689, with 8.1 per cent), base metal and metal articles (chapters 7273, with 6.7 per cent), and optical, measuring and precision instruments (chapters 90-92, with 5.2 per cent). On the import side, the majority of products that ASEAN sells to the EU are also machinery and electrical appliances (84-85), accounting for 43.8 per cent of the EU total imports from ASEAN in 2006. Other prominent commodities imported by the EU from ASEAN are chemicals (chapters 28-38, 8.8 per cent), textiles and apparel (chapters 50-63, 7.5 per cent), and plastics and plastic products (chapters 39-40, 5.5 per cent). In 2006, EU-25 merchandise trade with ASEAN showed a deficit of more than €29 billion, a steady increase since 2003. This deficit was greatest for office and telecommunications equipment (17.02 billion), agricultural products (-7.08 billion) as well as for textiles and clothing (-5. 68 billion). The increasing EU trade deficit in chemicals and the relatively comfortable EU position in the area of power and non electrical machinery, and to a lesser extent in transport equipment, should be particularly noted. In the area of services, trade with ASEAN was near balance in 2003, after two years showing a small deficit, but it eventually shifted to a surplus in 2004 and 2005. Bearing in mind that the EU has a net positive balance in world trade in services whereas ASEAN has a deficit with the rest of the world in this sector, the current situation is more in line with the two partners’ performances. The EU trade position is particularly comfortable in knowledge-intensive services such as insurance, financial services, computer and information services, and other business services; it has a negative trade balance in air transport, other transport, travel, and communication services. FDI: Traditionally FDI inflows to ASEAN were evenly split between three sources: Europe (EU and Switzerland), Asia (i.e. Japan, the Republic of Korea, Hong Kong and Taiwan, also including intraASEAN) and the rest of the world (primarily the United States).9 9
Since the total for the rest of the world includes offshore affiliates of OECD firms located in Bermuda, the Cayman Islands and elsewhere,
212
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
Over the past ten years or so, the EU has been the largest source of inward FDI into the ASEAN region, with more than US$ 13,000 million in 2006. Figure 1: FDI into ASEAN by source country/region 25000
mill. US$
20000 15000 10000 5000
19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05
0
Japan
Asian NIEs
US
Europe
Source: ASEAN Secretariat.
The geographical distribution of FDI inflows varies greatly by source country. The only constant is the importance of Singapore: the City-state absorbed 58 per cent of total European investment in ASEAN (compared to 61 per cent for American firms and 35 per cent for Japanese investors).10 Apart from Singapore, over the period 1995-2004, the other four main recipients of EU FDI are Malaysia (14 per cent), Brunei Darussalam (9 per cent), Indonesia (6 per cent) and Thailand (5 per cent). The other ASEAN member countries are only minor recipients in relative terms.11 Yet, in contrast to Japan and the US, the EU tends to be more actively involved in poorer ASEAN countries (Vietnam and Myanmar in particular).
10
11
the share of investment from this group would be lower if the ultimate owner could be determined. In cumulated terms over the period 1995-2004 (ASEAN Statistical Yearbook). Vietnam accounts for 4 per cent, followed by Myanmar (3 per cent), and the Philippines (2 per cent).
Assessing the Complementarity between the EU and ASEAN Economies
213
Although EU-15 is a net outward direct investor worldwide on the basis of stock figures, its liabilities are greater than its assets in the metal & mechanical industry (which encompasses office machinery and computers), confirming the weakness of the EU position in this area (computers). The bulk of the FDI from EU countries still goes into the manufacturing sector, but FDI has tended to switch into the financial and mining industries since 2001. European firms are dominant investors in oil producing countries such as Brunei, Indonesia and, to a lesser extent, Myanmar. In terms of the major recipients of EU FDI, the patterns of inward FDI vary across countries. The majority of FDI flows into the manufacturing and financial sectors of Singapore (with a possible rerouting effect), while in the case of Brunei, almost all EU FDI was drawn into the mining and quarrying industry, reflecting the resource rich nature of the country. The case of Thailand is more diverse, for the pattern of FDI has switched across the manufacturing, trade & commerce, and financial sectors over the years. In addition, there were net outflows from Thailand in the mining industry (in 2000 and 2001), in trade and commerce (2001-2003), financial services (2002-2003), and also in the manufacturing industry as a whole (2002-2003). Net outflows from Indonesia and Malaysia were also observed in the manufacturing sector in the early 2000s, while financial services attracted large EU FDI inflows, probably as a result of the financial crisis. Figure 2: FDI flows between the EU and ASEA, 1995-2005 US$ Million
10000 8000 6000 4000 2000 0
EU inflows
Source: Eurostat.
EU outflows
05 20
04 20
03 20
02 20
01 20
00 20
99 19
98 19
97 19
96 19
19
95
-2000
214
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
Finally, the image of a one way flow of investment between Europe and ASEAN which characterized earlier decades is no longer relevant. The EU is still a net direct investor, but the pattern of net flows has started being reversed (Figure 2). ASEAN firms are also beginning to invest worldwide, including in the EU. One of the most significant developments concerning ASEAN outward direct investment in recent years has been the rise of firms from Singapore and Malaysia as international investors, with Singapore alone being the fourteenth greatest outward investor world-wide in 2004, ahead of all but a handful of EU member countries. Among large ASEAN investors abroad, one should mention Singtel Ltd (Singapore) and Petronas (Malaysia) operating in the telecommunications and petroleum industries respectively. The Philippines and Thailand are also home of several large MNEs, with the former accounting for large investments in the EU in 2004. Together with Indonesia, they accounted for much smaller investments in the EU in 2005, while the rest of the ASEAN countries withdrew capital from the EU. Much of the ASEAN investment in Europe involves the acquisition of European firms, sometimes by ASEAN firms, which are still state-owned or controlled. In early 2006, the Singapore stateowned port operator, PSA International, made an unsuccessful £3.9 billion bid for P&O of the UK. In the late 1990s, Proton of Malaysia acquired the Lotus Car Group of the UK in search of advanced engine technology. The state investment agency in Singapore, Temasek Holdings, controls US$5 billion in assets in Europe. The motives for ASEAN investment in Europe range from portfolio diversification to the acquisition of technology, while market access does not appear to be a chief concern. III.
Assessing complementarity between ASEAN and the EU
In the area of economic integration theory, post-Vinerian thinking, which encompasses inter alia new trade and new growth theories elements, suggests that structural adjustment costs, implied by trade liberalization, should be a core element of the theory (Tongzon, 2005). In line with this thinking, we posit that the greater the complementarity between two economies, i.e. the more dissimilar their productive structures and specialization patterns, the higher the scope for trade expansion between them and the lower the costs
Assessing the Complementarity between the EU and ASEAN Economies
215
induced by resource reallocation. This in turn increases the chances for a more successful FTA.12 Structural similarity can be appraised through a variety of ways. This can be done by comparing and contrasting the productive and trade specialization of both regions, i.e. by delving into an analysis of revealed comparative advantage (RCA) and intra-industry trade (IIT) indices. A priori, the more diverse the revealed comparative advantages and the lower the IIT indices, the greater the level of complementarity between the two regions. Finally, we will refine and conclude the analysis by computing a trade complementarity (TC) index. A. Specialization, revealed comparative advantage and intraindustry trade At a highly aggregated level of analysis, the examination of productive specialization patterns reveals a clear complementarity between the two regions, with the EU being more engaged in services (representing more than 70 per cent of total EU-25 value added), whereas agriculture and mining are relatively important sectors for ASEAN. In the case of Myanmar, Laos and Cambodia, agriculture represented more than 55, 50 and 40 per cent, respectively, of total national value added in the early 2000s (UN figures). Given the abundance of its natural resources, mining and quarrying are important activities for Brunei (nearly a third of its value added) and also, albeit to a lesser extent, for Indonesia, Malaysia and Vietnam.13 Services, excluding wholesale & retail trade storage and transport, are an important sector only for Singapore, accounting for 65 per cent of its GDP in 2004 (World Development Indicators 2006). Moreover, in agriculture, there is some degree of complementarity in so far as ASEAN countries are not really engaged in dairy products, meat products and cereals (except rice), all products that 12
13
Note that, in contrast, Jacob Viner’s static Customs Union theory argues that trade creation effects are more likely, the more structurally similar the member states. Beyond its many restrictive assumptions, the theory totally ignores the issue of resource allocation costs induced by structural change. The manufacturing sector is the biggest in value added terms for Thailand, Malaysia and also to some extent for the Philippines and Indonesia, and it has been growing rapidly in the case of Vietnam since the launch of economic reforms.
216
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
are sensitive from the point of view of the EU. For the EU as a whole, four major groups of products account for nearly two thirds of the total agricultural production: meat, milk, fruits & vegetables and cereals. Relative differences in trade specialization between any two countries are illustrated by the index of revealed comparative advantage (RCA), which gives an indication of the relative importance in exports of a specific product or service to a country. The RCA index is computed as: x m RCA = §¨ i ·¸ §¨ i ·¸ , © X¹ © M¹ where xi and mi are exports and imports in product i from source to destination country, and where X and M are total exports and imports of the country. The higher the index, the higher is the revealed comparative advantage of the country. Negative figures denote a revealed comparative disadvantage. The benefit of an RCA index defined as above is that it takes into account both supply and demand side effects. ASEAN and the EU are used here as the two receiving and sending countries.14 According to ASEAN data, a first appraisal of RCA analysis indicates that the ASEAN strengths on EU markets are in textiles and apparel, machinery and electrical appliances, as well as in fats and oils. It is worth noting that the pattern of specialization has changed over time. As can be seen from Table 3, the revealed comparative advantage of ASEAN in textiles and apparel has decreased over time. When disaggregating by country (Table 4), the analysis of RCAs reveals for example that Brunei, Cambodia and Thailand have a clear revealed comparative advantage with the EU in textiles and apparel products; . The same holds true for Indonesia, although the trade structure of this country is more diversified, with mineral products and footwear being other areas of specialization. Malaysia has a slight RCA in fats and oil, and plastics; whereas the Philippines and Singapore exhibit a clear RCA in machinery and electrical appliances.
14
Therefore, this is a partial analysis of the RCA (of the EU for example).
Assessing the Complementarity between the EU and ASEAN Economies
217
Table 3: Revealed Comparative Advantage (RCA) in trade between the EU and ASEAN* Products (HS 1992) (A).ASEAN
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
1. Fats and oils (15) 0.156 0.162 2. Textiles and apparel (50-63) 0.124 0.121 3. Machinery appliances (84-85) -0.005 0.014
0.136
0.156
0.135
0.146
0.134
0.108
0.130
0.140
0.139
0.148
0.143
0.137
0.128
0.131
0.137
0.104
0.043
0.054
0.049
0.028
0.032
0.025
0.025
0.022
(B).EU-15 1. Chemicals (28-38) 2. Base metal (72-83) 3. Machinery appliances (84-85) 4.Vehicles (86-89)
0.015
0.014
0.006
0.006
0.006
0.004
0.004
0.001
0.004
0.007
0.016
0.015
0.005
0.006
0.006
0.008
0.008
0.008
0.007
0.007
-0.011 -0.018 -0.026 -0.027 -0.026 -0.019 -0.021 -0.025 -0.025 -0.023 0.018 0.016 0.006 0.001 0.002 0.001 0.001 0.003 0.003 0.005
Source: Author’s calculations based on the UN Comtrade statistics. *The data excludes Lao PDR., Myanmar and Vietnam. The data does not include Brunei in 2004-2005, and Cambodia in 1996-1999 and 2005.
The EU strengths on the ASEAN market are in vehicles, chemicals, and in base metals. As in the case of ASEAN, the RCA indices also imply that the pattern of EU specialization has changed over time. For all three most important product categories (chemicals, vehicles and machinery appliances), the value of the RCA index has either gradually declined and/or turned into the negative. It is worth noting at this juncture that the EU has been witnessing a relative comparative disadvantage in machinery and electrical appliances for a number of years, but not with all ASEAN member-countries. Further calculations at the 4-digit level confirm these findings (Table 5). ASEAN has a relative comparative advantage in some areas of food and agricultural products, such as coffee (code 0901) and palm oil (1511), as well as in textile, clothing and footwear such as jerseys (code 6110) and footwear with outer soles of leather (code 6403). The region is also relatively engaged in the production of natural rubber (code 4001), as well as coal, briquettes and similar
218
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
solid fuels (2701). As suggested above, ASEAN also has a clear advantage in some areas of the medium to high tech sector; this is the case for automatic data processing machines and units thereof (code 8471), for transmission apparatus for radio-telephony (8525), and for motor vehicles for the transport of goods (8704). In these medium to high-tech industries, the EU still has nevertheless a clear comparative advantage in a larger number of industry segments, compared with ASEAN. This is for example the case for: turbojets (8411), machines having individual functions (8479), motor cars (8703) and powered aircraft (8802). Table 4: Revealed Comparative Advantages (RCAs) of Individual ASEAN Countries with the EU, 1994 - 2003 Brunei Darussalam Textiles and apparel Cambodia Textiles and apparel Indonesia Mineral products Textiles and apparel Footwear Chemicals Machinery and electrical appliances Malaysia Fats and oils Plastics Chemicals Vehicles Myanmar Textiles Machinery and electrical appliances Philippines Textiles and apparel Machinery and electrical appliances Chemicals Singapore Machinery and electrical a. Thailand Textiles and apparel Prepared foodstuffs Chemicals
1994
1995
2000
2002
2003
0.03
0.03
0.02
0.62
0.11
n.a.
n.a.
0.88
0.88
0.83
0.02 0.30 0.09 -0.11 -0.46
004 0.23 0.10 -0.13 -0.42
0.07 0.20 0.07 -0.18 -0.15
0.09 0.17 0.04 -0.19 -0.22
0.11 0.17 0.05 -0.19 -0.21
0.04 0.05 -0.08
0.05 0.07 -0.04 -0.13
0.02 0.02 -0.07 -0.01
0.03 0.04 -0.04 -0.03
0.06 0.05 -0.06 -0.04
n.a. n.a.
n.a. n.a.
0.56 -0.21
0.66 -0.29
0.72 -0.16
0.17 -0.11
0.11 -0.03
0.03 0.17
0.02 0.21
0.03 0.15
-0.15
-0.15
-0.11
-0.12
-0.13
0.37
0.36
0.24
0.22
0.18
0.16 0.04 -0.12
0.08 0.03 -0.11
0.08 0.07 -0.14
0.07 0.03 -0.14
0.06 0.04 -0.16
Source: Authors’ calculations, based on data from the ASEAN Secretariat.
Assessing the Complementarity between the EU and ASEAN Economies
219
Table 5: Revealed Comparative Advantage Indices (RCA) EUASEAN (4-digit level, 2004 and 2005) (to be cont.) HS 2004 2005 Product name PREPARED FOODSTUFFS 0901 -0.0058 -0.0069 Coffee 1511 -0.0166 -0.0163 Palm oil (excl. chemically modified) Coconut "copra", palm kernel or babassu oil 1513 -0.0085 -0.0090 and fractions thereof Undenatured ethyl alcohol of an alcoholic 2208 0.0083 0.0091 strength of < 80%; spirits, liqueurs and other spirituous beverages Coal; briquettes, ovoids and similar solid fuels 2701 -0.0087 -0.0107 manufactured from coal Petroleum oils and oils obtained from 2710 0.0073 0.0047 bituminous minerals (excl. crude) CHEMICALS 2922 -0.0160 -0.0276 Oxygen-function amino-compounds 2930 -0.0155 -0.0195 Organo-sulphur compounds 2935 -0.0061 -0.0072 Sulphonamides Medicines consisting of mixed or unmixed 3004 0.0192 0.0187 products for therapeutic or prophylactic uses Mixtures of odoriferous substances and 3302 0.0047 0.0042 mixtures, incl. alcoholic solutions. Beauty or make-up prepar. and preparations 3304 0.0045 0.0051 for the care of the skin, incl. sunscreen Industrial monocarboxylic fatty acids; acid 3823 -0.0047 -0.0045 oils from refining; industrial fatty alcohol PLASTICS 4001 -0.0168 -0.0165 Natural rubber Articles of apparel and clothing accessories, 4015 -0.0065 -0.0067 incl. gloves, mittens and mitts, of vulcanized rubber 4418 -0.0037 -0.0042 Builders' joinery and carpentry CLOTHING AND FOOTWEAR T-shirts, singlets and other vests, knitted or 6109 -0.0057 -0.0048 crocheted Jerseys, pullovers, cardigans, waistcoats, 6110 -0.0164 -0.0129 knitted or crocheted Men's or boys' suits, jackets, blazers, trousers, 6203 -0.0051 -0.0044 shorts.. Women's or girls' suits, jackets, blazers, 6204 -0.0062 -0.0045 dresses, skirts, trousers… Footwear with outer soles and uppers of 6402 -0.0117 -0.0088 rubber or plastics
220
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
Table 5 (cont.): Revealed Comparative Advantage Indices (RCA) EU-ASEAN (4-digit level, 2004 and 2005) HS
2004
2005
Product name Footwear with outer soles of rubber, 6403 -0.0248 -0.0238 plastics, leather or composition leather and uppers of leather Footwear with outer soles of rubber, 6404 -0.0081 -0.0069 plastics, leather or composition leather and uppers of textile materials MISCELLANEOUS 7102 0.0055 0.0055 Diamonds Articles of jewellery and parts thereof 7113 -0.0054 -0.0063 (excl. articles > 100 years old 7219 0.0049 0.0051 Flat-rolled products of stainless steel Tubes, pipes and hollow profiles, seamless, 7304 0.0029 0.0051 of iron or steel (excl. products of cast iron) Tubes and pipes, having circular cross7305 0.0028 0.0062 sections and an external diameter of > 406,4 mm MACHINERY AND ELECTRICAL APPLIANCES Parts suitable for use solely or principally 8409 0.0052 0.0054 with internal combustion piston engine of heading 8407 or 8408 Turbojets, turbopropellers and other gas 8411 0.0178 0.0158 turbines 8413 0.0056 0.0055 Pumps for liquids Air or vacuum pumps; air or other gas 8414 0.0062 0.0048 compressors and fans; 8415 -0.0111 -0.0098 Air conditioning Machinery, plant or laboratory equipment whether or not electrically heated for the 8419 0.0057 0.0051 treatment of materials by a process involving a change of temperature 8421 0.0050 0.0049 Centrifuges 8422 0.0075 0.0089 Dishwashing machines Parts suitable for use solely or principally 8431 0.0048 0.0051 with the machinery of heading 8425 to 8430, n.e.s. 8443 0.0051 0.0048 Printing machinery Automatic data-processing machines and 8471 -0.1006 -0.0854 units thereof; magnetic or optical readers Parts and accessories suitable for use solely 8473 -0.0313 -0.0409 or principally with machines of heading 8469 to 8472, n.e.s.
Assessing the Complementarity between the EU and ASEAN Economies
221
Table 5 (cont.): Revealed Comparative Advantage Indices (RCA) EU-ASEAN (4-digit level, 2004 and 2005) HS 8477
2004 2005 Product name 0.0055 0.0062 Machinery for working rubber or plastics Machines and mechanical appliances 8479 0.0141 0.0122 having individual functions Taps, cocks, valves and similar appliances 8481 0.0068 0.0073 for pipes, boiler shells, tanks…. Ball or roller bearings (excl. steel balls of 8482 0.0049 0.0056 heading 7326); parts thereof Transmission shafts, incl. camshafts and 8483 0.0044 0.0044 crankshafts, and cranks 8521 -0.0068 -0.0058 Video recording and reproduction apparatus Transmission apparatus for radio-telephony, 8525 -0.0192 -0.0143 radio-telegraphy, radio-broadcasting or television Reception apparatus for radio-telephony, 8527 -0.0081 -0.0067 radio-telegraphy or radio-broadcasting Electrical apparatus for switching or 8536 0.0041 0.0042 protecting electrical circuits Diodes, transistors and similar 8541 0.0101 0.0086 semiconductor devices VEHICLES Motor cars and other motor vehicles 8703 0.0169 0.0161 principally designed for the transport of persons Motor vehicles for the transport of goods, 8704 -0.0071 -0.0083 incl. chassis with engine and cab Parts and accessories for tractors, motor 8708 0.0070 0.0075 vehicles for the transport of ten or more persons 8802 0.0274 0.0384 Powered aircraft; spacecraft, incl. satellites Parts of aircraft and spacecraft of heading 8803 0.0042 0.0044 8801 or 8802, n.e.s. OPTICAL, PRECISION & MUSICAL INSTRUMENTS Furniture and parts thereof, n.e.s. (excl. 9403 -0.0144 -0.0144 seats and medical, surgical, dental or veterinary furniture) Note: The calculations are based on the Harmonized System Codes (HS Code), at the four digit level, and the results have to be interpreted from the perspective of the EU, i.e. in the expression (xi/X) – (mi/M), X are exports by the EU to the ASEAN. Only products from the industries with relatively high RCA indices are reported.
222
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
It follows that trade specialization of the two regions, at an industry level, does not follow a clear cut pattern. Although the traditional industries (i.e. wood & wood products, and textiles, apparel and footwear) are still relatively important for most ASEAN economies, it cannot be said that the EU as a whole has disengaged itself from these industries. Owing to its heterogeneous composition, the EU has indeed a trade advantage in high-technology high value added products (such as scientific instruments) with most ASEAN countries, but it is still relatively engaged in some sunset industries, such as textiles. This is to be related to the EU’s heterogeneity, with the new EU Member Countries and some of its ‘old’ member countries such as Italy and Portugal, being still relatively specialized in these traditional industries. With regard to EU-ASEAN intra-industry trade, a preliminary appraisal at the two-digit level of analysis and using a standard Grubel-Lloyd index15 shows that the trade structure between the EU and ASEAN is apparently rather dissimilar with only 38 out of the 99 chapters exhibiting a GL index larger than 50 in 2003. Beyond this aggregate picture, however, the situation varies widely across countries. The trade structure of the EU on the one hand and some individual ASEAN member countries on the other suggests that their trade follows a Ricardian type of pattern. This implies a possible degree of complementarity between the two partners. Grubel-Lloyd (GL) indices are indeed low in the case of Brunei (12.69 in 2003), Cambodia (27.70) and Myanmar (21.41).16 By contrast, these indices are relatively high for countries such as Malaysia (68.5 in 2003, up from 46.2 in 1993), Thailand (52.7 in 2003), or Singapore (51.6 in 2003). The use of sectoral IIT indices leads to additional and more detailed conclusions. In particular there seems to be still ample scope for complementarity even between countries with less dissimilar levels of development. In the case of Malaysia for instance, the IIT index is indeed quite low in chemicals (23), textiles and 15
This is computed as follows: Bj
16
ª § X j M j ·º «1 ¨¨ ¸¸ » x100 , with ¬« © X j M j ¹ »¼
0 d B
j
d 100
where, Xj = exports
of industry j and Mj = imports of industry j. For an analysis in terms of dynamic GL indices, see Andreosso-O’Callaghan et al. (2006). Authors’ calculations based on ASEAN Secretariat data.
Assessing the Complementarity between the EU and ASEAN Economies
223
apparel (26) and vehicles (36). In contrast, it is high in machinery and electrical equipment (86), where more competition can be expected. The same observation holds true in the case of Thailand, and to a lesser extent for the Philippines, Singapore and Indonesia (Table 6). The relatively high level of aggregation (2-digit level) may, however, account to a large extent for these results and complementarity may still prevail if trade in machinery and electrical equipment for instance involves trade in parts and components. Table 6: Grubel-Lloyd indices between the EU and individual ASEAN member countries (2-digit level), 2003 Chemicals Textiles Machinery Transportation and allied and electrical products equipment (28 – 38) (50 – 63) (84 – 85) (86 – 89) 48.0 8.0 57.0 50.0 Indonesia 23.0 26.0 86.0 36.0 Malaysia 12.0 10.0 58.0 37.0 Philippines 24.0 34.0 58.0 84.0 Singapore 20.0 22.0 75.0 59.9 Thailand Source: Authors’ calculations based on trade data from the ASEAN Secretariat.
Further calculations at a finer degree of disaggregation suggest that there is indeed a fair degree of complementarity between the EU and ASEAN as a whole. Using data at the 4-digit level of analysis, a total of 52 products are found to be intensively traded between the two regions. These products, which account for 60 per cent of total bilateral trade, are broken down as follows: three in the vegetable products category (06-15), one in foodstuffs (16-24), three in mineral products (25-27), six in chemicals and allied products (28-38), three in plastics and rubber (39-40), one in wood and wood products (44-49), two in textiles (50-63), three in footwear and headgear (64-67), two in glass and stone (68-71), eighteen in machinery/electrical (84-85), five in transportation (8689), and five in miscellaneous activities (90-97). Out of these 52 intensively traded products, 17 products only record high GL indices (above 50). These 17 products (shown in Table 7) are primarily concentrated in machinery and electrical appliances, in transportation, and in miscellaneous products. A closer look at these products suggests that they may give rise to substantial trade in parts and components.
Assessing the Complementarity between the EU and ASEAN Economies
227
Two major conclusions can be drawn from this exercise: first, for the large majority of products traded between the two regions, IIT ratios are relatively low, denoting important trade complementarities between the two regions. Second, high GL indices are often observed in trade in parts and components or in products where segmentation of production is likely to prevail,17 reflecting further complementarities, although of a different kind. B. Using a trade complementarity index This preliminary appraisal of the issue of structural complementarity through intra-industry trade can be complemented by resorting to a trade complementarity index as defined in Vaillant and Ons (2002). This index helps measuring the exact intensity of trade complementarity between two regions. Trade complementarity of country A in relation to country B and in industry s is computed as follows: s TCI AB
x As m Bs . s s t wA t wA
where x As is the share of industry s in country A’s exports, and s is the share of industry s in world imports, from country A’s t wA
viewpoint. This index, which is in fact the product of the export specialization index of country A and the import specialization index of country B, measures the level of similarity between the export supply of a country and the import demand of one of its partners. A value greater than unity implies the existence of a strong complementarity between the export specialization of country A and the import specialization of its partner B. Because of the unavailability of some of the data necessary to calculate the index, the analysis in this section is based only on a two digit level of analysis. Computing TCI indices between each of the ASEAN countries and the EU at the two digit level gives the results shown in Table 8 (only results that are above unity are reported).
17
This is the case for parts and accessories in the automobile industry (8708), or for parts of electrical products (8529, 8534, 8541, 8542).
228
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
Table 8: Trade complementarity between the EU and each individual ASEAN country (2005) Country
Brunei Darussalam*
Other Transport Equipment (79) Articles of apparel and clothing (84)
Trade complement arity index 2.21 1.62
Indonesia
Cork and wood (63) Paper, paperboard articles (64) Textile yarn (65) Articles of apparel and clothing (84)
6.47 2.53 1.91 2.39
Malaysia
Cork and wood (63) Office machines & automatic data equipment (75) Telecommunications & sound recording apparatus (76) Electrical machinery, apparatus & appliances n.e.s. (77) Office machines & automatic data equipment (75) Electrical machinery, apparatus & appliances n.e.s. (77) Articles of apparel and clothing (84)
3.35 3.26 1.47 1.85
The Philippines
Singapore
Thailand
Vietnam**
Products (SITC Rev.3)
3.81 3.49 2.22
Organic chemicals (51) Office machines & automatic data equipment (75) Telecommunications & sound recording apparatus (76) Electrical machinery, apparatus & appliances n.e.s. (77) Rubber manufactures (62) Office machines & automatic data equipment (75) Telecommunications & sound recording apparatus (76) Articles of apparel and clothing (84)
2.34 2.88 1.27 2.29
Articles of apparel and clothing (84)
3.60
2.50 2.07 0.96 1.47
Source: Authors’ calculations based on data from the UN Comtrade. (*, **) Figure for 2003 and 2002 respectively.
Assessing the Complementarity between the EU and ASEAN Economies
229
These results show that trade complementarity between the EU and individual ASEAN countries exists both in the case of laborintensive industries (apparel and clothing) and of medium to hightech industries. All ASEAN countries in the table, except Singapore and Malaysia, are complementary with the EU as a whole in the area of apparel and clothing, implying that imports of these products by the EU can be met by these ASEAN countries given their export specialization patterns, ceteris paribus.18 These data also reveals that the productive weakness of the EU in office machines & automatic data equipment (at the two digit level) can be compensated for by exports from Malaysia, the Philippines, Singapore and Thailand. Other medium to high-tech areas where there is trade complementarity between the EU and ASEAN countries, in favor of the latter, are telecommunication equipment (with Malaysia, Singapore and Thailand), electrical machinery (Malaysia, Philippines and Singapore), and photographic apparatus, optical goods and watches with Singapore. IV.
Conclusions
Despite earlier efforts aimed at enhancing trade linkages between the EU and ASEAN, the latter is still not regarded as a crucial economic (trade and investment) partner for the EU, whereas the EU has a relatively important economic presence in ASEAN countries. One way to revive and rebalance the partnership between the two regions is through the signing of a FTA agreement, a top issue at the moment. Indeed, the new EU Strategy on Global Competitiveness, which aims at boosting EU competitiveness through the development of privileged economic partnerships with selected emerging economies, has identified ASEAN as one such potential partner. The chapter argues that the success of the planned FTA depends on the degree of complementarity between the two regions, with complementarity mirroring the structural (i.e. productive and trade) dissimilarity between the two regions. The appraisal of complementarity through the use of several indices, leads to the following conclusions. First, on a sectoral level, 18
The ceteris paribus clause is here important since the extent to which these ASEAN countries can, after the FTA, benefit from trade expansion in the EU, depends on the supply characteristics of other clothing producers (in particular, China and India).
230
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
there is a clear manufacturing-services complementarity, with ASEAN (in particular ASEAN-6, excluding Singapore) being relatively involved in manufacturing activities and the EU in services, in particular in knowledge-based services, such as financial services and business services. Second, at the manufacturing level, both ASEAN and EU countries display an extremely diverse production structure and differing productivity levels, with each region in toto being relatively engaged in sunset as well as in sunrise industries. However, the relative comparative advantage of ASEAN in automatic data processing machines (an EU structural weakness) should be noted. Third, for the large majority of manufactured products traded between the two regions, IIT ratios are relatively low. This suggests that inter-industry trade still looms large, denoting important trade complementarities based on Ricardian comparative advantages. Fourth, the analysis at the four digit level further shows that the majority of products with high IIT ratios are concentrated in industries where trade in parts and components can be expected to be high. This suggests that complementarity between the two regions may also be based on fragmentation of production. Fifth, the use of a complementarity index, calculated at the twodigit level, shows that Malaysia, the Philippines, Singapore and Thailand, four economies that are structurally close to the EU average, are nevertheless complementary with the EU in office machines and automatic data equipment. Telecommunications equipment, electrical machinery as well as photographic apparatus, optical goods and watches are other industries where a number of ASEAN countries are complementary with the EU, implying that the FTA would stimulate imports of these products into the EU. Evidence of a multi-faceted trade complementarity between the two regions renders the FTA all the more desirable. Moreover, the deepening of the partnership may also contribute to reduce the trade and investment asymmetry existing between the two regions.
Assessing the Complementarity between the EU and ASEAN Economies
231
References
Ando, M., Kimura F. (2007), Fragmentation in East Asia – Further Evidence, mimeo, Hitotsubashi University, Tokyo. Ando, M., Kimura, F. (2003), The formation of international production and distribution networks in East Asia, NBER Working paper No. 0167, Washington, D.C. Andreosso-O’Callaghan, B., Low, L., Nicolas, F., Petschiri, A., Thomsen, S., Uprasen, U. (2006), A Qualitative Analysis of a Potential Free Trade Agreement between the EU and ASEAN, Commissioned Report for the European Commission, DG Trade, Brussels, 28th April. ASEAN Secretariat (2006), ASEAN Secretariat Data Bases, Jakarta. Athukorala, P-C., Yamashita, N. (2005), Production Fragmentation and Trade Integration: East Asia in a Global Context, Mimeo, Australian National University. Athukorala, P., Yamashita, N. (2006), Production fragmentation and trade integration: East Asia in a global context, The North American Journal of Economics and Finance 17(3), 233-256. Borrus, M., Dieter, E., Haggard, S. (2000), Networks in Asia: Rivalry or Riches?, Routledge, London and New York. CEC (Commission of the European Communities) (1980), The ECASEAN Cooperation Agreement, Official Journal L 144, 10th June, Brussels. Clarete, R.L. (2005), Philippines: Ex-post Effects of Trade Liberalization in the Philippines, paper for United Nations Conference on Trade and Development, in Geneva, Switzerland. EUROSTAT (Various Years) Statistical data bases, Luxembourg. Hew, D. (2005), Roadmap to an ASEAN Economic Community, Singapore Institute of Southeast Asian Studies.
232
Bernadette Andreosso-O’Callaghan and Françoise Nicolas
Liang, M. (2005), Singapore’s Trade Policies, Priorities and Options, ASEAN Economic Bulletin 22 (1), 49-59. Matthews, J. A., Dong-Sung, C. (2000), Tiger Technology: The Creation of a Semiconductor Industry in East Asia, Cambridge University Press, Cambridge. McKendrick, D., Doner, G., Richard, F., Haggard, S. (2000), From Silicon Valley to Singapore: Location and Competitive Advantage in the Hard Disk Drive Industry, Stanford University Press, Stanford. Ng, F., Yeats, A. (2003), Major Trade Trends in East Asia: What are Their Implications for Regional Cooperation and Growth?, Policy Research Working Paper 3084, World Bank, Washington DC. Soesastro, H. (2005), ASEAN Economic Community: Concept, Costs, and Benefits, in: Hew, D. (ed.), Roadmap to an ASEAN Economic Community, Singapore Institute of Southeast Asian Studies, Singapore. Strange, R., Slater, J., Molteni, C. (2000), The European Union and ASEAN: Trade and Investment Issues, Macmillan Press, London. Tongzon, J.L. (2005), Trade Policy in the Philippines - Treading a Cautious Path, ASEAN Economic Bulletin 22 (1), 35–48. Vaillant, M., Ons, A. (2002), Preferential Trading Arrangements between the European Union and South America: the Political Economy of Free Trade Zones in practice, The World Economy 25 (10), 1433-68. WTO (1998), Report of the Working Group on the Relationship between Trade and Investment to the General Council, WT/WGTI/2. WTO (2006), World Trade Report, Geneva.
Assessing the Complementarity between the EU and ASEAN Economies
Appendix
HS Codes: 01-05 06-15 16-24 25-27 28-38 39-40 41-43 44-49 50-63 64-67 68-71 72-83 84-85 86-89 90-97 98-99
Animal & Animal Products Vegetable Products Foodstuffs Mineral Products Chemicals & Allied Industries Plastics / Rubbers Raw Hides, Skins, Leather, & Furs Wood & Wood Products Textiles Footwear / Headgear Stone / Glass Metals Machinery / Electrical Transportation Miscellaneous Service
233
Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh*
Vietnam’s Export to the EU: An Overview and Assessment using the Constant Market Share based Approach Abstract In order to deepen the understanding of Vietnam’s export to the EU market, this paper uses the constant market share (CMS) approach to investigate the influence of separate crucial factors determining Vietnam’s export performance. The major data source is the COMTRADE database, from which import data of reporting countries during the period 1997-2004 are compiled at 2-digit SITC product categories and grouped for 8 major trading partners. Our empirical findings suggest that Vietnam’s strong export performance is attributed to a favourable world trade effect and a positive competitiveness residual. However, the negative contribution of commodity composition and market distribution effects reveals that there is an ample scope for improving its trade pattern through product and market diversification. I. Introduction Over the last 20 years of reforms, Vietnam has achieved great successes in terms of Gross Domestic Product (GDP) growth, macroeconomic stabilization, export expansion, Foreign Direct Investment (FDI) attraction and poverty reduction. The economy grew at an impressive level of 7.5 per cent per year during 1990-2004, while GDP per capita increased by 5.7 per cent per year. As a result of fast economic development, the poverty rate sharply fell from 70 per cent in 1986 to 37 per cent in 1998 and 19.5 per cent in 2004. *
This research paper has benefited from the European Commission’s financial support along the EU-Vietnam Small Projects Facility in Economic Cooperation. It is part of the study “Vietnam’s export to the EU: market access and facilitation” implemented by the Department for Trade Policy and International Economic Integration Studies of the CIEM (Ministry for Planning and Investment, Vietnam).
236 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
Among the crucial factors contributing to these remarkable achievements is the export sector, which constitutes more than 50 per cent of GDP with an annual growth rate of 16 per cent during the last ten years. The impressive performance of the export sector is attributed to a great deal of continuing efforts of improving trade policies toward a sound regime of trade liberalization through strengthened international economic integration. In 1995, Vietnam joined the Association of Southeast Asian Nations (ASEAN) and committed itself to fulfill the agreements under the ASEAN Free Trade Area (AFTA) by 2006. In the same year, Vietnam applied for WTO membership and became the 150th WTO member on 11 January 2007. The country is also member of the APEC since 1998. In 2000, Vietnam and the United States (US) signed a Bilateral Trade Agreement (BTA) which became effective in December 2001. Since then, Vietnam has also joined several regional integration clubs such as the ASEAN-China Free Trade Area (ACFTA) in 2002, ASEAN-Japan Comprehensive Economic Partnership in 2003, and ASEAN-Korea Free Trade Area in 2006. With 25 member countries, the European Union (EU) is currently the largest economic pillar in the world with approximately 20 per cent of world GDP in 2005. It is the world's leading exporter and the second-largest importer of goods, and accounts for 25 per cent of world trading and service activities. The EU has always been an important partner of Vietnam. This close relationship has developed rapidly and vigorously since Vietnam and the EU established diplomatic relations in 1990 and signed a trade agreement in 1992. By 2005, the total two-way trade value reached 8,108 million USD, of which Vietnam’s exports to the EU valued at 5,520 million USD, equivalent to 17 per cent of Vietnam’s total exports. Although the export value increased year by year, the share of the EU market in Vietnam’s export has declined from about 22 per cent during 1998-99 to 17 per cent in 2005. Meanwhile, the shares of Vietnam’s exports to other major trading partners like the US have increased. This phenomenon raised great concerns about Vietnam’s export competitiveness, especially on the EU market. In parallel, thanks to a strong involvement in the international division of labour, China’s foreign trade has developed rapidly since the implementation of reforms and the opening-up policies in 1978. While trade policy reforms before 1987 were considered to be at an experimental stage, the Central government facilitated from 1988 on the expan-
Vietnam’s Export to the EU
237
sion of multinational corporation activities and cross-border production networks that used FDI and non-FDI relationships to exploit various advantages of China’s location (low costs of labor, economies of scale, spatial concentration). In addition, China became a full member of the WTO on 11 December 2001. As a large developing country, this event has exerted and will continue to exert profound influence on the world economy. China has grown into one of the main competitors in international markets. It has become the third largest exporter in the world in 2004 and is expected to become the first largest by the beginning of the next decade (Gaulier et al., 2006). This phenomenon is important for the rest of the world, and especially for the less developing exporting countries like Vietnam. It requires efforts to find out what are the determinants of Vietnam’s export performance regarding the world market in general, and the EU market in particular. At the specific case of Vietnam, some studies attempted to determine whether China’s emergence has diverted foreign trade from its Asian neighbour. However, most of them restricted their analysis to revealed comparative advantage calculation or only used qualitative methods to assess Vietnam’s export performance. To our acknowledgement, only one quantitative analysis attempted to measure Vietnam’s export competitiveness during 1988-1996 (Vo, 1998). However, the level of Vietnam’s international integration at that period was limited, so that the quantitative findings are somewhat unconvincing in explaining impacts of trade policy reforms on export growth, not to mention the difficulty to access reliable data. Accordingly, this paper presents a research attempt to overcome such limitations by using CMS analysis to investigate the influence of separate crucial factors determining Vietnam’s export performance. The major data source used for our analysis is the COMTRADE database, where import data provided by reporting countries will be compiled by 2-digit SITC product category and grouped for 8 major trading partners of Vietnam during the period 1997-2004. Other data, mostly used for qualitative assessment, will be obtained from the General Statistics Office (GSO) and other reference books. The rest of the paper is organized as follows. Section II presents an overview of Vietnam’s export: it covers a brief review of trade policy reforms in Vietnam and an overall description of its exports, with a focus on the EU market. Section III then quantitatively
238 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
analyzes factors determining Vietnam’s export performance by using the CMS approach. Finally, section IV concludes the paper and summarizes our main research findings. II. An overview of Vietnam’s export A. Trade reforms in Vietnam since Doi Moi Although the announcement of Doi Moi (“Renewal” in Vietnamese) inaugurated the transition period in 1986, trade reforms in Vietnam were, however, only initiated in 1989 and have experienced a lot of ups and downs so far. The trade reform process was strongly supported between 1989 and the early 1990s, suffering a standstill during the mid-1990s, resuming with a convincing speed since 1998. As pointed out by Vo (2005), trade reforms that are closely interconnected with the export performance in Vietnam largely refer to trading rights, import protection, and export promotion with incentives, which are summarized in Table 1. Evidently over the last 15 years of trade reforms, there has been a significant liberalization of the foreign trade regime in Vietnam. Non-tariff import restrictions which create trade distortions have been abolished gradually, while favourable mechanisms have been applied to exports. Also, the success of the export sector is not without the government’s prompt responses towards trade in determined moves for joining the WTO.1 A great harmonization in legal framework and policies has been made for consistency with international norms and practices. Trade promotion activities become more diversified; management mechanism of export and import policies has been more flexible and predictable. At the same time, Vietnamese entrepreneurs and traders have gradually become more experienced, professional and dynamic, resulting in an improvement and expansion of production, business and trade capacity (CIEM, 2006).
1
Formerly, Vietnam’s tariff schedule was composed of three tariff rates: Most Favoured Nation (MFN), preferential and normal tariff rates (the latter were usually 50 per cent higher than the MFN ones). Vietnam’s simple MFN average was higher than the preferential one, but lower than that of neighbouring countries such as Thailand, Philippines and Indonesia (Nguyen, 2002). As a consequence of the recent accession to the WTO and the implementation of AFTA, the MFN and preferential tariff schedules replaced the much higher previous tariff rates.
Vietnam’s Export to the EU
239
Table 1: Major developments and reforms of Vietnam’s trade policy (to be continued) 1. Trading rights -1989 State monopoly in foreign trade dominant. 1989- Entry conditions for participating in trading activities relaxed but 1997 still heavily restrictive. 1998 Abolishment of trade licenses as most significant step toward trade liberalization. 1999 Free trade (both import and export) of commodities. 2002 Foreign owned enterprises granted right to export goods other than those they produce. 2. Import protection Non-Tariff Barriers (NTBs): Quantitative restrictions and foreign exchange management 1998 Introduction of extremely strict control over foreign exchange. Saving account in foreign currency permitted. Partial (80 per cent) foreign exchange surrender requirements imposed on enterprises having foreign exchange accounts. 1994- Introduction and increased use of quantitative controls on goods 2000 during 1994-2000, up from 5 in 1996 to 12 items in 2000. 2000 Balancing requirement of foreign exchange for foreign owned enterprises relaxed, allowing them to buy foreign currency from domestic banks to repay loans provided by offshore banks. 2001 A turning point in trade reform: a trade policy roadmap publicly issued the first time, allowing for prediction of export-import environment over the next 5 years (2001-05). Quantitative restrictions reduced to two products (petroleum and sugar). 2003 Introduction of tariff-rate quotas, imposed on agricultural products such as cotton, tobacco materials, and salt. Tariff barriers 1988 Introduction of simplified tariff scheme: tariff rates of 0-60 per cent imposed on 130 product categories. 1989 Maximum tariff rates increased to 120 per cent on some luxury goods. Tariff coverage reduced to 80 product categories.
240 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
Table 1 (cont.): Major developments and reforms of Vietnam’s trade policy 1991 Law on Import and Export Duties approved, separating normal tariffs from preferential ones. 1992 HS system of tariff nomenclature introduced with detailed and consolidated tariff schedule. 2002 Tariffs reduced significantly to 16 per cent on average. 60 per cent of all tariffs below 10 per cent. 2005 Tariffs increased to about 18 per cent due to tariffication of NTBs in place of quantitative restrictions. 3. Export promotion 1991 Restrictions on imports relaxed by granting duty free access to inputs imported for export production. Implemented with the establishment of export processing zones and the introduction of import duty rebate scheme. 1998 Export duties basically removed, leaving only two commodities (crude oil and scrap metal) subject to export duties. 2004 The VAT, initiated in 1999 to replace turnover tax, significantly improved with a reduction from four tax lines to three tax lines, allowing export firms to enjoy the same VAT tax band for imports as firms producing for domestic consumption. Unification of corporate income tax to 28 per cent for both industries and services. Under unified system, concessionary rate applied to tax profits from export production, subjected to the degree of exportoriented production. Source: Vo (2005) and Athukorala (2005).
Nevertheless, the current Vietnam’s trade policy is characterized by a ‘dual’ approach that the country has been pursuing for a number of years. On the one hand, Vietnam states its intention to promote export-oriented sectors, where export manufacturing firms, especially the Foreign Owned Enterprises (FIEs) with regional and global competitiveness, play the leading role. Thus, trade reform has helped the private sector by enhancing its access to imported inputs and to export outlets. By 1999, it accounted for nearly threequarters of all trading firms and half of export and imports (Nguyen, 2002). However, considerable bias against exports still exists in several sub-sectors where Vietnam has an ample scope of achieving export success, such as garments, plastic products, leather goods, ceramics and other manufacturing. The anti-export bias in
Vietnam’s Export to the EU
241
the incentive structure is creating a major constraint for a sound development of the private sector, especially with regard to small and medium-sized enterprises – which are considered to be a driving force for export expansion (Athukorala, 2005). In addition, exportsupporting measures are still weak and insufficient. Infrastructure for export activities lacks both in quantity and quality; information provision for analysis and forecast of import source and export markets is not timely and sufficient. The linkage between domestic enterprises in trade activities are still weak, failing to take advantages of industry and trade associations for better trade promotion (CIEM, 2006). On the other hand, the foreign trade regime is still considered to be restrictive to protect specific industries such as steel, cement and fertilizer. The Vietnamese government maintains protection of capital intensive State-Owned Enterprises (SOEs) which are dominant in import-substituting sectors. Although they benefit from FDI inflows (FIEs establish joint-ventures with those SOEs to enjoy import-substitution status), they remain generally weak and inefficient. Such issues reflect Vietnam’s rather special circumstances as an Asian country in transition both from a centralized planned economy to a market economy, and from a developing economy to an industrialized one. Indeed, the SOEs enjoy rents from existing barriers to trade because of their leading role in policy making, revenue collection and distribution in the absence of macroeconomic and budget management, or market institutions. To remove these irrationalities, Vietnam needs to conduct trade reforms in parallel with a wide range of macroeconomic and structural reforms of SOEs, State Owned Commercial Banks (SOCBs), and tax administration (Auffret, 2003; Vo, 2005). B. A view on Vietnam’s exports Vietnam shows an outstanding performance in achieving global, as well as regional economic integration. It currently has wide trade relations with more than 220 nations and territories. During the period 1993 to 2004, Vietnam’s trade openness (exports plus imports of goods in per cent of GDP) rose from 54.2 per cent to 123.9 per cent, while the market share of its exports (exports of goods in percent of world imports of goods) more than tribbled, from 0.08 per cent to 0.29 per cent (IMF, 2005). This is an impressive performance record compared to other countries in East Asia: China in particular, was less open to trade (26.9 per cent in 1993
242 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
and 68.2 per cent in 2004) and ‘only’ doubled its export market share (from 5.74 per cent to 9.64 per cent during the same period). Moreover, China, India and Vietnam were the only Asian countries which strengthened their positions in world imports between 1993 and 2003. The trade liberalization measures implemented under various Regional Trade Agreements (RTAs) during the 1990s, together with the region’s rapid economic growth, have led to significant growth of Vietnam’s trade with its East Asian neighbours (Figure 1). This phenomenon also suggests the ongoing participation of Vietnam in regional production sharing and outsourcing networks that China has contributed to intensify. As Asian trade becomes more and more centred on China, the latter has been overtaking Japan as the major trading partner in the region. However, Vietnam is heavily dependent on final demand outside the region, as domestic consumption still remains a relatively weak component of demand in most Asian economies. Thanks to the VN-US BTA effective since late 2001, Vietnam has dramatically increased its exports to the US market, from around 1 billion USD in 2001 to nearly 6 billion USD in 2005 (GSO, 2005). Following the implementation of the BTA, the US is currently among the largest importers of Vietnam, together with the EU, Japan, China and ASEAN4 (see Table 2). In recent years, growth of Vietnam’s exports to almost all major destinations accelerated, but these rates differed among importing regions. Notably, while Vietnam’s exports to the EU market just increased modestly from 15 per cent per annum during 1997-2001 to 18 per cent per annum during 2001-2004, its export growth to the US market quickly went up from 29 per cent per annum to over 70 per cent per annum during the corresponding periods (see Table 3). Table 2: Vietnam’s export structure by destination (period average, in per cent) ASEAN4 CHINA NICs JAPAN OCEANIA EU15 US ROW 19978 6 11 21 7 29 7 11 2001 20017 8 10 16 7 27 17 8 2004 Source: Authors’ calculations from COMTRADE statistics. Note: ASEAN4: Indonesia, Malaysia, Philippines, Thailand. NICs: Newly Industrializing Countries of East Asia (Singapore, Hong Kong, South Korea).
Vietnam’s Export to the EU
243
Figure 1: Development of Vietnam’s exports by major trading partners 35,000 ROW
30,000
China
m illio n U S D
25,000
Japan
20,000 15,000
US
10,000
ASEAN
5,000
EU-15
0 1998
1999
2000
2001
2002
2003
2004
2005
Source: GSO.
Table 3: Annual growth of Vietnam’s exports by destination (in per cent) ASEAN4 CHINA NICs JAPAN OCEANIA EU15 19972001 11 30 9 4 25 15 20012004 18 35 19 14 19 18 19972004 14 32 13 8 22 16 Source: Authors’ calculations from COMTRADE statistics.
US ROW 29
3
71
18
46
9
The impressive export development of Vietnam during recent years can be attributed to several factors such as growth of world trade, higher world market prices and market expansion. According to COMTRADE statistics, world trade growth accelerated from 3 per cent per annum during 1997-2001 to 13 per cent per annum during 2001-2004. Vietnam benefited from higher world market prices in a number of leading export goods like crude oil, rice, coffee, rubber and coal (World Bank, 2005). According to CIEM (2006), the average growth of 11.5 per cent in export prices led to the increase of the export value by 3.3 billion USD. The largest rise in export prices was registered for crude oil (40.7 per cent), coffee (24.7 per cent), coal (20.7 per cent), rubber (17.9 per cent), tea (15.9 per cent), rice (14.5 per cent) and cashew nuts (12.5 per cent).
244 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
Meanwhile, the export volume increased by about 9.4 per cent on average, resulting in a climb of about 2.4 billion USD in export turnover. The items with the greatest increase in export volume were coal (53.8 per cent), rice (28.1 per cent), peanuts (26.1 per cent), and rubber (11.9 per cent). Table 4: Vietnam’s export structure by commodity (in per cent) Commodity group (one-digit SITC) Primary products (SITC 0-4) 0 - Food and live animals 1 - Beverages and tobacco 2 - Crude materials, inedible, except fuels 3 - Mineral fuels, lubricants and related materials 4 - Animal and vegetable oils, fats and waxes Manufactured products (SITC 5-8) 5 - Chemicals and related products, n.e.s. 6 - Manufactured goods classified chiefly by material 7 - Machinery and transport equipment 8 - Miscellaneous manufactured articles 9 - Commodities and transactions not classified elsewhere in the SITC Total
1997
2000
2001
2002
2003
2004
47.60
50.20
47.60
49.62
46.64
40.56
25.62
20.13
20.26
24.65
22.00
15.93
0.14
0.14
0.14
0.45
0.79
0.13
2.57
2.44
2.39
3.09
3.13
2.42
18.98
27.30
24.72
21.36
20.60
22.04
0.29
0.18
0.08
0.07
0.11
0.04
52.40
49.80
52.40
50.38
53.36
59.44
1.00
0.92
1.24
1.57
1.69
1.14
5.24
5.56
5.59
6.73
6.72
6.13
6.80
6.68
8.31
8.00
8.90
9.34
39.18
36.44
36.98
34.06
36.03
42.52
0.19 100.0
0.20 100.0
0.28 100.0
0.01 100.0
0.02 100.0
0.31 100.0
Source: Authors’ calculations from COMTRADE and GSO statistics.
Vietnam’s Export to the EU
245
With an apparent increase of manufactured products from 52 per cent in 1997 to 60 per cent in 2004 at the expense of primary products, Vietnam´s exports structure improved, in accordance with the export-led growth strategy (see Table 4). The trade structure clearly has diversified, highlighted by a notable increase of SITC-8 (miscellaneous manufactured articles) from 39.18 per cent to 42.52 per cent. Like other Asian countries, Vietnam’s external trade becomes more and more driven by its involvement in the regional segmentation of production processes, as shown by the rise of SITC-7 (which includes parts and components) from 6.8 per cent to 9.34 per cent. However, more than two-thirds of export revenues are still concentrated on seven key commodities - namely crude oil, textile and garment, footwear, seafood, agricultural products, furniture, and electronics appliances. Looking at the export structure in more detail at the 2-digit SITC aggregates, we find the following major export items in 2004: crude oil (about 20 per cent of total exports), garments (16.5 per cent), footwear (16 per cent), agricultural and fishery products (11 per cent and 8 per cent respectively). In short, the export sector is still dominated by natural resource intensive, labour-intensive and low value-added products; many of them are also vulnerable to fluctuations in the world markets. Although Vietnam should boost the existing leading export sectors, China’s comparative advantage in the same products has increased global competition and may affect Vietnam’s ability to improve its position in the world. On the other side, such international specialization raises concerns about the appropriate export pattern for the country to pursue its export-led growth strategy in the long run. C. Vietnam’s exports to the EU Economic and trade relations between Vietnam and the EU have been established since long time. This special relationship has developed rapidly and vigorously since Vietnam and EU established diplomatic relations in 1990. The EU is one of Vietnam’s largest trading partner and export market, absorbing nearly one-fifth of Vietnam’s exports during the past ten years. Two-way trade has increased four times over the last ten years, and was worth 8.1 billion USD in 2005 (EUECC, 2006). Generally speaking, the EU grants Vietnam a special status like other developing countries. It applies a trade regime with relatively liberal regulations towards imports from Third World countries.
246 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
Import licensing is generally not required for commodities originating from those countries, except for some sensitive products such as agricultural goods, tobacco, weapons and products under surveillance, quantitative restrictions, or safeguard measures. In addition, the EU has applied the Generalized System of Preferences (GSP) since 1971, defining preferential arrangements for duties on EU’s imports originating from Third World countries. Under the GSP, imports from least developed and developing nations are not subject to normal customs duties, and traders in the EU enjoy incentives when importing products from developing countries. The latest version of the EU’s GSP2 was approved on 27 June 2005, effective for 3-years from January 1, 2006 to December 31, 2008. Accordingly, Vietnam is currently benefiting from the GSP. Regarding the commodity structure (see Table 5), Vietnam’s exports to the EU are heavily concentrated in a limited range of commodities, almost without changes over the past years. In particular, the bulk of Vietnam’s export to the EU (more than 80 per cent) consists of manufactured products, notably SITC-8 (almost three-quarters). On the opposite, Vietnam’s main imports from the EU are machine tools, electrical equipment, and pharmaceutical products to meet Vietnam’s demand for modern technology and materials as the industrial development proceeds. On the 2-digit SITC level, four export sectors can be identified as potential interesting items in our investigation (see Table 6). On the one hand, footwear and garments are manufacturing activities where Vietnam is confronted with China’s competition on the EU market and, at the same time, with important barriers to access EU markets.3 On the other hand, fishery and agricultural products are among commodities exported to the EU market where Vietnamese producers have to meet quality standards and very strict rules4 on the grounds of food safety. 2 3
4
Council Regulation (EC) No 980/2005. According to Gaulier et al. (2006), the rise of China in European imports more than compensated the relative decline of other Asian economies. Since 2003, the EU has a new Common Fisheries Policy (CFP), aiming at “achieving biologically, environmentally, and economically sustainable fisheries by, inter alia, better conserving fish stocks, protecting the marine environment, and ensuring the economic viability of the European fleets” (WTO, 2004).
Vietnam’s Export to the EU
247
Table 5: Vietnam’s export structure with the EU by commodity per cent of total exports to EU Commodity group 1997 2000 2001 2002 2003 2004 (one-digit SITC) Primary products 16.04 13.28 12.16 14.05 16.28 13.52 (0-4) 0 - Food and live 14.34 11.38 10.22 11.88 14.02 11.33 animals 1 - Beverages and 0.01 0.04 0.03 0.08 0.05 0.03 tobacco 2 - Crude materials, 1.27 0.87 0.81 1.52 1.71 1.41 inedible, except fuels 3 - Mineral fuels, lubricants and related 0.41 0.98 1.09 0.57 0.50 0.75 materials 4 - Animal and 0.01 0.01 0.01 0.01 0.00 0.00 vegetable oils, fats and waxes Manufactured 83.96 86.72 87.84 85.95 83.72 86.48 products (5-8) 5 - Chemicals and 0.74 0.23 0.20 0.37 0.27 0.18 related products, n.e.s. 6 - Manufactured 4.49 7.17 7.55 8.30 8.20 7.73 goods classified chiefly by material 7 - Machinery and 1.45 2.75 3.13 5.33 6.23 5.01 transport equipment 8 - Miscellaneous 77.26 76.55 76.93 71.94 68.83 73.47 manufactured articles 9 - Commodities and 0.02 0.02 0.03 0.00 0.19 0.09 transactions not classified elsewhere in the SITC 100.0 100.0 100.0 100.0 100.0 100.0 Total Source: Authors’ calculations from COMTRADE and GSO statistics.
248 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
Table 6: Shares of Vietnam's major export commodities to the EU per cent of total exports to EU
Commodities (two-digit SITC)
1997 2
2000 2
SITC-03: Fish, crustaceans, molluscs and aquatic invertebrates, and preparations thereof 11 7 SITC-07: Coffee, tea, cocoa, spices, and manufactures thereof 3 5 SITC-82: Furniture, and parts thereof; bedding, mattresses, mattress supports, cushions and similar stuffed furnishings 5 4 SITC-83: Travel goods, handbags and similar containers 25 19 SITC-84: Articles of apparel and clothing accessories 40 44 SITC-85: Footwear 4 4 SITC-89: Miscellaneous manufactured articles, n.e.s. Source: Authors’ calculations from COMTRADE statistics.
2001 3
2004 4
6
6
5
9
4
3
18
13
45 4
44 4
All of these export products are subject to surveillance and safeguard measures. 1. Vietnam is the EU’s second largest footwear supplier after China. The majority of Vietnam’s footwear output (90 per cent) is for export, concentrating on sport shoes (approximately 70 per cent) and lady shoes (20 per cent) (ActionAid and Lefaso, 2006). In terms of turnover, Vietnam’s total export of footwear products increased by approximately 24.6 per cent a year during 1995-2005. However, the share of footwear exports to the EU has declined from 74.5 per cent in 2001 to 59 per cent in 2005 (EUECC, 2006). Since 2006, the sector also has to arrange with the EU anti-dumping petition that imposed progressive tariffs on Vietnamese leatherupper shoes. The anti-dumping proceedings has almost instantly affected footwear enterprises with a sharp fall in orders and production of 50 per cent, causing production contraction, economic losses with income reduction of 15–60 per cent and massive layoffs. Some firms changed to produce imitation leather products, but they find it difficult to compete with China due to its abundant materials and frequent changes in shoes design/styles. 2. In the last ten years, Vietnam witnessed a remarkable performance of the garment and textile industry, with an increase of
Vietnam’s Export to the EU
249
export revenues by 21 per cent on average per year (EUECC, 2006). The three major markets are the US, with 53.6 per cent of Vietnam's garment and textile export value, followed by the EU and Japan (VENO, 2005). The signing of the Vietnam-US Bilateral Trade Agreement resulted in spectacular growth for Vietnamese garment and textile exports. On the opposite, its export to the EU grew much more slowly - around 10 per cent in 2005 (despite the removal of quotas on textiles and clothing by 31 December 2004), compared to the average annual growth of 21 per cent during 20002005. One of the reasons behind lower garment and textile exports was the strong competition from China. According to VENO (2005), output of Vietnam’s garments and textiles during the first five months of 2005 accounted for just about 1.3 per cent market share of the EU market, against 26 per cent for China. 3. As one of the leading foreign currency earners of Vietnam, the export of fishery products maintains a high growth rate, with an average annual increase of 16.5 per cent for seafood during 19952005 (EUECC, 2006). In recent years, Vietnam’s exports to the US increased sharply, although still suffering from anti-dumping tariffs imposed since 2003 by the US on two major fishery products (catfish and black tiger shrimp). Vietnam found its way out of this difficulty by changing the export commodity structure and exploring new international markets. Ranked as the third largest customer of Vietnam’s fishery (behind Japan and US), the EU market maintained a high growth rate in 2005, by 79 per cent. The rapid development of export markets has encouraged new seafood processing enterprises, increasing from 405 in 2004 to 439 in 2005. Among them, by late 2005, a total of 171 enterprises with access to the EU market and 300 enterprises to the US market (EUECC, 2006). 4. Agricultural products are currently exported to hundreds of countries worldwide and contribute substantially to Vietnam’s total export revenues. In addition to policy reforms in agriculture that have induced incentives for farmers, it is market and export orientation that plays a key role in the sector development, resulting in an average growth of 4.5 per cent per annum during 2000-2005 (EUECC, 2006). By region, Asia is the largest market for Vietnam’s agricultural products, mainly rice, rubber, fruits and vegetables, pepper and cashews. The emerging US market consumes rather coffee, pepper and pineapple juice, while the EU market imports coffee, honey, processed fruits and vegetables
250 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
(EUECC, 2006). With current low import tariffs imposed on agrobased products, the EU market is viewed to have a high potential for the entrance of Vietnamese agro-products, provided that they comply with the EU’s strict quality and preservative requirements. In addition, agricultural products continue to be subject to safeguard clauses applied under the Common Agricultural Policy. III.
Vietnam’s export performance: a Constant Market Share Approach
A. The Constant Market Share Model Constant Market Share (CMS) analysis is a methodology which was applied for the first time to international trade flows by Tyszynski (1951). It was originally used in the empirical studies of structural change in industrial and regional economics. Basically, the CMS model is built on the assumption that without changes abroad and maintained competitiveness at home, a country’s export share in the world market should remain unchanged over time. Any deviation of the actual export growth from this constant market share is broken down into four components: world trade, commodity composition, market distribution, and a residual usually referred to as an ex-post competitiveness. The decomposition method has to be applied to export data at a disaggregated level with respect to markets and products. It then allows the analysis of the separate influence of these different factors on export performance. Following Leamer and Stern (1970), the CMS identity decomposes actual change in a country’s exports between two periods as follows: V' V
rV ¦ ri r Vi ¦¦ rij ri Vij ¦¦ Vij' Vij rijVij N 1
i
i
j
V
3
4
X’s export value of commodity i in period 2 value of X’s exports to country j in period 1
V
V’
2
Where: value of country X’s total exports in period 1 value of country X’s total exports in period 2 X’s export value of commodity i in period 1
V V’ V V’
i i j i j
j
ij
value of X’s exports to country j in period 2 X’s export value of commodity i to country j in period 1
Vietnam’s Export to the EU
V’ r r r
i
ij
ij
251
X’s export value of commodity i to country j in period 2 percentage increase in total world exports from period 1 to 2 percentage increase in world exports of commodity i from period 1 to 2 percentage increase in world exports of commodity i to
country j from period 1 to 2 Appendix describes the decomposition procedures for producing the aforementioned CMS identity. Country X’s export growth can then be attributed to: 1. Overall trade effect. It represents what the country’s growth in exports would have been if they had grown at the same rate as the total value of world exports, i.e. if the country maintained its market share. 2. Commodity composition effect. It indicates the difference in the growth rate of exports that each commodity may have experienced compared to the world average. A country’s export may grow because it concentrates in commodities with growth rates higher than the world average. 3. Market distribution effect. A country’s export may grow because it exports to geographical markets experiencing high growth. 4. Competitiveness effect. It indicates the difference between the actual increase in exports and the hypothetical increase if the country maintained its export share in each market, with respect to each commodity group and to each country. A positive sign of the residual reveals an improvement in export competitiveness. In spite of some conceptual and empirical limitations as indicated by Richardson (1971), the CMS approach is a commonly accepted procedure of accounting for the sources of a country’s export growth, depending on the availability of data. Bowen and Pelzman (1984), or Azam and Azam (1994), used a CMS model to analyze the export performance of the US in manufactured goods. In empirical studies for East Asian countries, the CMS approach was applied by Chen et al. (1999) who evaluated the ex-post competitiveness of China’s exports in agri-food products during 198096. Indeed, China was a net exporter of agri-food to the world market in the period under consideration. The authors attempted to examine whether the rise of China’s export in agri-food products was associated with a rise in its international competitiveness. The CMS results revealed a downward trend of China’s competitiveness in agri-food products over the years 1986 to 1996. Hence, the ob-
252 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
served increase in agri-food exports was mainly due to growth of world import demand and a favourable market distribution effect. Similarly, Juswanto and Mulyanti (2003) examined Indonesia’s manufactured exports during the 1990s. The analysis revealed that Indonesia’s export growth had a severe problem with its commodity composition and low response to world demand. The ADB Institute (2002) also used CMS estimates as standard export performance indicators to determine whether the emerging East Asian economies lost export competitiveness before the crisis in mid-1997. Contrary to the belief that Asia was no longer a dominant force, the study found that the East Asian economies registered high export growth in the early 1990s. For Vietnam, Vo (1998) used the CMS approach to analyze the export performance of the country during 1988-96. During that period however, despite a strong pace of trade reforms, Vietnam just started its international integration, so that exports were not actually seen as a leverage of the Vietnamese economy. B. Data and methodology As pointed by Richardson (1971), better results are obtained by using disaggregated data in such calculations. Consequently, the CMS decomposition is estimated here by compiling data from COMTRADE on an annual basis at a two-digit SITC level of commodity aggregation, where import data provided by reporting countries are used as proxies for export values. The highly aggregate nature of the commodity data used here is due to statistical constraints. Vietnam’s world trading partners are grouped into eight destinations: ASEAN4 (Indonesia, Malaysia, Philippines, Thailand), CHINA, EU-155 (Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, The Netherlands, Portugal, Spain, Sweden, United Kingdom), JAPAN, NICs6 (newly industrialized Asian countries Singapore, Hong Kong, South Korea), OCEANIA (Australia, New Zealand), US, and ROW (Rest Of the World, i.e. remaining countries). 5
6
During the period under consideration (1997-2004), the share of Vietnam’s exports to the ten new members of the EU was not significant compared to EU-15, not to mention the accessibility of data. Thus, EU-15 can be used as a good proxy for the EU as a whole. Taiwan is not included due to data unavailability and thus is grouped in ROW.
Vietnam’s Export to the EU
253
Our analysis focuses on the period from 1997 to 2004, considering data availability for Vietnam and its major trading partners. The CMS decomposition of Vietnam’s exports for the period under consideration allows us to analyze the performance of Vietnam’s exports over the past ten years. However, in an attempt to link export performance with major influential events or reforms during the past ten years, the CMS analysis is carried out at the same time for the two consecutive sub-periods 1997-2001 and 2001-2004. The year 2001 is chosen as a breaking point in order to capture the impact of the Vietnam-US Bilateral Trade Agreement (effective since December 2001), which has been thought to accelerate Vietnam’s exports to the US significantly. Correspondingly, the year 2001 also benchmarks China’s entry into the WTO. Because of similar commitments required by the WTO, the Chinese experience should facilitate trade prospects for Vietnam after joining the WTO. Therefore, the same CMS-based analysis is also conducted for China under the two sub-periods 1997-2001 and 2001-2004 in order to assess the impact of China’s accession to the WTO on its export performance. C. Decomposition results As illustrated in Table 7, CMS decomposition for the period 19972004 as a whole shows that the rise in Vietnam’s exports can be largely explained by a world trade effect (which contributed about 30 per cent to the overall export performance) and a competitiveness effect (which accounted for 82.5 per cent), whereas commodity and market effects are found to have a negative contribution. Indeed, COMTRADE statistics indicate that world trade growth has accelerated dramatically during the period under consideration, leading to a significant rise in export volume and price. The competitiveness effect, on the other hand, attests the impressive progress Vietnam has made towards reforming its trade system since the early 1990s. However, decomposition for the two sub-periods reveals a significant difference in the shares of determining factors. During the first sub-period 1997-2001, the competitiveness effect dominated Vietnam’s export growth, explaining up to 95 per cent of rising exports, whereas the world trade effect contributed to only 24 per cent. This development attests the considerable speed of the trade reform process following the regional crisis in mid-1997. By contrast, the second sub-period 2001-2004 was marked by a
254 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
strong growth of world import demand (from 3 per cent per annum during 1997-2001 to 13 per cent per annum during 2001-2004), resulting in a rapidly increased contribution of this factor to Vietnam’s export growth. Overall, despite ‘adverse’ effects of commodity and market distribution, a significant difference between actual export growth and the world trade effect during both subperiods and the whole period is clearly attributable to a ‘competitiveness’ residual effect. After its WTO accession, China’s export growth increased sharply, from an average annual level of 10 per cent during 19972001 to 23 per cent during the period 2001-2004. The two most important factors contributing to rising exports are world trade and competitiveness effects, with no substantial change between the two sub-periods (see T. 8). However, Vietnam stands out for its impressive competitiveness effect on the whole compared to China, albeit more moderately during 2001-2004. The difference in the results may be explained by the fact that Vietnam has liberalized its trade regime to a considerably greater extent than China. In comparison, China’s export competitiveness less eroded than Vietnam´s in the second sub-period 2001-2004 due to China´s WTO entry and its consequences (market access, business facilitation, etc.). The postWTO accession period also witnessed a positive contribution of the commodity composition effect (changing from -17 per cent to 2 per cent). This suggests that China has significantly improved its product composition by concentrating its exports on commodities in which demand has been more dynamic than the world average. Table 7: CMS decomposition of Vietnam's export performance during 1997-2004 Category
1997-2001 2001-2004 1997-2004 USD Value Share USD Value Share USD Value Share in % in % in % Total change 5,260,619,750 100 13,365,044,283 100 18,625,664,033 100 in exp. value World trade 1,256,836,258 24 6,262,342,354 47 5,732,922,336 30.8 effect Commodity -422,051,335 -8 -461,205,131 -3 -1,007,568,445 -5.4 composition effect Market distri- -579,982,065 -11 -1,311,451,241 -10 -1,458,758,528 -7.8 bution effect Competitive- 5,005,816,892 95 8,875,358,300 66 15,359,068,670 82.5 ness effect
Source: Authors’ calculations from COMTRADE statistics.
Vietnam’s Export to the EU
255
Table 8: CMS decomposition of China's export performance during 1997-2004 Category
1997-2001 2001- 2004 1997-2004 USD Value Share USD Value Share USD Value Share in per in per in per cent cent cent 133,967,049,014 100 358,843,450,617 100 492,810,499,631 100
Total change in export value World trade 60,889,322,458 effect Commodity -22,828,346,829 composition effect Market -16,992,817,016 distribution effect Competi112,898,890,401 tiveness effect
45
145,478,848,897
41
180,451,164,127
37
-17
6,836,214,078
2
-25,977,701,172
-5
-13
-56,351,535,734
-16
-37,871,181,297
-8
84
262,879,923,377
73
376,208,217,974
76
Source: Authors’ calculations from COMTRADE statistics.
China’s case provides strong evidence that WTO membership will influence Vietnam’s export growth through a product composition effect. Further discussion should therefore be given to Vietnam’s export structure by commodity. This might also provide explanation for the country’s declining export share to the EU in recent years. The negative contribution of the commodity effect to Vietnam’s export growth implies that Vietnam had concentrated on commodities in which the market was growing relatively slower than the world average. An analysis of COMTRADE data shows that except crude oil, the majority of Vietnam’s leading export commodities under SITC 03, SITC 07, SITC 84, and SITC 85 registered growth of import demand at a lower rate than the average level of world demand. However, the positive move towards trade diversification in accordance with the export-led growth strategy (as shown in Table 3), is consistent with the findings of the CMSbased analysis, where the commodity composition effect in the subperiod 2001-2004, despite still negative, was less severe than the previous one. To some extent, a limited export diversification (or concentration on few miscellaneous articles) may be among the key reasons explaining Vietnam’s eroding export capacity to the EU in comparison with other large markets. This may help to explain the modest fall of Vietnam’s export share with the EU compared to
256 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
Vietnam’s export share with the whole world. As a consequence, more attention should be paid to increased export diversification so that Vietnam may improve its access to the EU market and, at the same time, mitigate negative competition pressures arisen from China. The CMS decomposition also points to the negative sign of the market distribution effect in Vietnam´s exports, that is even larger than the product composition effect. China registered the same negative effect (even worse), although its exports increased sharply during the period under consideration. It implies that WTO accession in 2001 did not improve the market distribution effect as the latter still contributed negatively (about 16 per cent) to the overall export performance. This result indicates that both countries concentrated their exports on markets with lower demand growth than the world average. As Vietnam’s trade flows with its East Asian neighbours rapidly increased over the last decade, the negative contribution of the market distribution effect was largely due to the collapse of growth throughout the region that began in mid-1997, as well as Japan’s recession during the period under consideration. Table 9 reports the average annual rates of import growth for the eighth major trading partners of Vietnam during the period 19972004. While the first sub-period was marked by the East Asian crisis, we can see that Japan, the US and the ROW registered a lower import growth than the world demand during the second subperiod. However, these markets accounted for 40 per cent of Vietnam’s exports. Moreover, the negative sign of the market distribution effect in the second sub-period 2001-2004 seems to contradict the expectation that the Vietnam-US Bilateral Trade Agreement would have a great impact on Vietnam’s export performance. Hence, the CMS analysis reveals that Vietnam and China had a severe problem with their market distribution. Table 9: Growth of imports from Vietnam in of selected regions Import growth from Vietnam in per cent period ASEAN4 CHINA NICs JAPAN OCEANIA EU15 US ROW World 1997-3 14 -2 1 -1 5 7 1 3 2001 200113 32 13 9 20 15 9 10 13 2004
Source: Authors’ calculations from COMTRADE statistics.
Vietnam’s Export to the EU
257
IV. Conclusions Since the early 1990s, Vietnam made an impressive progress in opening its trade system. Its trade flows increased rapidly over the past decade due to Vietnam´s participation in bilateral and regional trade agreements. In addition, the comprehensive increase in its export market shares attest the successful implementation of its outward-oriented growth strategy. In order to deepen understanding our of Vietnam’s export performance, this paper uses the CMS approach to investigate the influence of separate crucial factors. The major data source is the COMTRADE database, from which import data of reporting countries during the period 1997-2004 are compiled by 2-digit SITC product categories and grouped for 8 major trading partners. The period under consideration is split into two sub-periods 1997-2001 and 2001-2004 in order to capture the impact of the Vietnam-US Bilateral Trade Agreement. The quantitative analysis is also conducted for China to compare the findings and to learn lessons from China´s post-WTO accession. The empirical results suggest that Vietnam’s strong export performance can be attributed to favourable factors. Among them is the rapid increase in world import demand. Over the period 19972004, the growth rate of world trade influenced Vietnam’s export performance to a large extent with a positive contribution in both sub-periods. However, despite ‘adverse’ effects of commodity and market distribution, a significant difference between the actual export growth and growth of world trade during both sub-periods and as well as in the whole period is clearly attributable to a residual competitiveness effect. This positive competitiveness effect is linked to the impressive progress of Vietnam in reforming its trade system since the early 1990s. The negative sign of commodity composition effect reveals that the majority of Vietnam’s exports remain resource-based and labour-intensive products and, in fact, world demand for those commodities tends to be low. In addition to maintaining current leading export sectors, the negative commodity composition effect may suggest that in a longer term perspective, Vietnam should think of expanding capital intensive, as well as more advanced endproducts. With regards to labour-intensive industries, especially those facing strong competition from China, Vietnam needs to raise labour productivity and to improve knowledge content in final products. Although the commodity composition effect was still
258 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
negative, we could see a sign of improvement in the export commodity structure in the sub-period 2001-2004. In this context, it should be noted that the level of diversification of Vietnam’s export to EU markets lags behind that of the country’s export to the world market as a whole, which may be one of the reasons explaining the decrease of Vietnam’s export share to the EU in recent years. Therefore, there exists a considerable potential for upgrading Vietnam’s export performance through product diversification. The negative contribution of the market distribution effect to Vietnam´s overall export performance suggests that there is an ample scope for improving market diversification beyond large markets like Japan, the EU and the US, which remain the main export markets for final goods. Market strengthening and expansion should take into account the important contributions of trade promotion and trade negotiations with the proactive supports of the government, in line with the international practice. The strong emergence of China as a huge export ‘machine’ gives a clear message, namely that the less developing countries may be marginalized and cannot compete with him unless they pay due attention to China-induced challenges. Thus, even though Vietnam recently achieved high growth rates, there is no guarantee that it will be able to sustain such high growth in the future, unless it improves its competitiveness as international integration proceeds. Opportunities of Vietnam’s WTO accession heavily depend on how the government succeeds in changing its industrial structure towards more value-added products.
References
ActionAid and Lefaso (2006), Half of A Million Vietnamese Footwear Jobs at risk: where is the balance between trade and development?, ActionAid International Vietnam and Vietnam Leather and Footwear Association, ActionAid Vietnam Publications, Hanoi. Asian Development Bank (ADB) Institute (2002), Did East-Asian Developing Economies Lose Export Competitiveness in the PreCrisis 1990s?, Research Paper Series No. 34, Asian Development Bank Institute, Tokyo.
Vietnam’s Export to the EU
259
Athukorala, P.C. (2006), Trade Policy Reforms and the Structure of Protection in Vietnam, World Economy 29 (2), 161-187. Auffret, R. (2003), Trade Reform in Vietnam: Opportunities with Emerging Challenges, World Bank Policy Research Working Paper 3076, World Bank, Washington DC. Azam, G., Azam, F. (1994), Sources of Change in US Manufactured Exports during the Eighties, Journal of Economics and Finance 18 (1), 31-42. Bowen, H.P., Pelzman, J. (1984), US Export Competitiveness: 1962-1977, Applied Economics 16, 461-473. Central Institute for Economic Management (CIEM) (2006), Vietnam’s Economy in 2005, the Publishing House of Political Theory, Hanoi. Chen, K., Xu, L., Duan, Y. (2000), Ex-post Competitiveness of China’s Export in Agric-food Products: 1980-96, Agribusiness 16 (3), 281-294. EU Economic and Commercial Counsellor (2006), 2006 Report on Vietnam, Hanoi. Gaulier, G., Lemoine, F., Unal-Kesenci, D. (2006), China’s emergence and the reorganisation of trade flows in Asia, CEPII Working Papers 2006-05. Juswanto, W., Mulyanti, P. (2003), Indonesia’s Manufactured Exports: A Constant Market Shares Analysis, Journal Keuangan dan Moneter 6(2), 97-106. Leamer, E.E., Stern, R. (1970), Quantitative International Economics, Aldine, Chicago. Nguyen, T.T. (2002), Vietnam’s international trade regime and comparative advantage, CAS Discussion Paper No. 37, Centre for ASEAN Studies/CIMDA. Richardson, J.D. (1971), Constant Market Shares Analysis of Export Growth, Journal of International Economics 1 (2), 227239. Tumbarello P. (2006), Does Vietnam overtrade with its neighboring countries? A regional investigation using gravity model, in: International Monetary Fund, Vietnam. Selected Issues, IMF Country Report n°06/20, International Monetary Fund: Washington DC.
260 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh
Tyszynski, H. (1951), World Trade in Manufactured Commodities, 1899-1950, The Manchester School 19(3), 272-304. Vietnam Economic News Online (VENO) (2005), A challenge for Textiles and Garments, Ministry of Trade, http://www.ven.org.vn/view_news.php?id=4817 (last visit: 25 October 2006. Vo, T. T. (1998), Determinants of Export growth: An Examination of Vietnam’s Export Competitiveness since 1988, Master of Development Economics (MDE) thesis, National Economics University, Hanoi, Vietnam. Vo, T.T., (2005), Vietnam’s Trade liberalization and International Economic Integration: Evolution, Problems and Challenges, Asian Economic Bulletin 22 (1), 75-94. World Bank (2005), Taking Stock: an Update on Vietnam’s Economic Developments and Reforms, Consultative Group Meeting for Vietnam, Hanoi, December 6-7. World Trade Organization (WTO) (2004), Trade Policy Review: European Communities, WT/TPR/S/136, Trade Policy Review Body, Geneva.
Appendix: The CMS model
The value of country X’s exports in period 1 is defined as: V = V = V =V
¦¦ i
j
ij
¦ i
i
¦ j
j
(A1)
Analyzing exports of country X at the first level, we view exports in a theoretical context of a single good and a single market. It is argued that if country X maintained its export share in the global market then exports would increase by rV, where r refers to the percentage increase in total world exports during the period under consideration. Thus, actual change in export values of country X can be expressed as: V’–V = rV + (V’ –V – rV) (A2) According to identity (A2), the change in export growth between period 1 and period 2 (V’ – V) is decomposed into a portion attrib-
Vietnam’s Export to the EU
261
utable to the overall growth in world exports (rV) and an unexplained ‘competitiveness’ residual (V’ – V – rV). At the two-level analysis, identity (A2) is decomposed further to analyze a more diverse set of exports with i commodities. Thus identity (A2) is now equal to:
¦ r V + ¦ (V’ – V - r V ) = ¦ (r – r + r )V + ¦ (V’ – V - r V ) = ¦ (rV ) + ¦ (r -r)V + ¦ (V’ – V - r V )
V’ – V =
i
i
i
i
i
i
i
i
i
i
i
i
V’ – V = (rV) +
i
i
i
i
i
i
i
i
i
i
i
i
i
i
¦ (r -r)V + ¦ (V’ – V - r V ) i
i
i
i
i
i
i
(A3)
i
(1) (2) (3) As shown in identity (A3), the changes in country X’s exports are divided into smaller parts, explaining changes due to (1) the general rise in world exports, (2) the commodity composition of country X’s export in period 1, and (3) an unexplained ‘competitiveness’ residual indicating the difference between the actual export increase and the hypothetical increase if X had maintained its export share with regard to each commodity group. The commodity composition ¦ (r -r)V implies that if world i
i
i
exports of commodity i grow by more than total world exports, then (r – r) will have a positive sign. The positive sign would indicate i
that country X’s exports were concentrated on commodities in which the market was growing relatively fast. Otherwise (r – r) i
would be negative if X’s exports were concentrated on the markets that were growing relatively slower than total world exports. In the real world, country X will export i commodities to j different markets. We come to the three-level analysis, where the CMS method will allow us to differentiate country X’s exports by i commodities and j destinations. At this level of analysis, changes in country X’s exports will now be decomposed further as:
262 Thi Anh-Dao Tran, Tri Thanh Vo, Hien Minh Dinh, Anh Duong Nguyen, Thien Hoang Pham and Quang Long Trinh V’ – V =
¦ ¦ V’ - ¦ ¦ V i
=
ij
j
j
ij
ij
i
j
ij
–V -r V ) ij
ij
ij
¦ ¦ (r – r + r – r + r )V + ¦ ¦ (V’ i
=
i
¦ ¦ r V + ¦ ¦ (V’ i
=
ij
j
i
j
¦¦ i
i
ij
ij
i
j
ij
–V –r V ) ij
ij
ij
i
ij
i
ij
ij
ij
¦ ¦ (V’
(rV – r V + r V – r V + r V ) +
j
ij
ij
i
j
ij
–V
ij
–r V ) ij
=
ij
¦ ¦ rV + ¦ ¦ (r - r)V + ¦ ¦ (r i
ij
j
i
i
j
ij
i
j
ij
– r )V + i
ij
¦ ¦ (V’ i
ij
j
–V –r V ) ij
ij
V’ – V = (rV) +
ij
¦ (r – r) V + ¦ ¦ (r i
i
i.
i
j
ij
– r) V + i
ij
¦ ¦ (V’ i
j
ij
–V
–r V ) ij
ij
(A4)
ij
(1) (2) (3) (4) The CMS decomposition by commodity i and destination j as shown in identity (A4) represents the analysis at level three, where the growth of country X’s aggregate exports is attributed to four components: (1) changes in world demand for exports; (2) the commodity composition of X’s export; (3) the market distribution of X’s export; and (4) an unexplained ‘competitiveness’ residual implying the difference between actual export increase and the hypothetical increase if X had maintained its share of exports of each commodity group to each country. Identity (A4) defines the market composition effect as
¦¦ i
j
(r – r ) V . This effect will be positive if country X concentrates its ij
i
ij
export in the markets that grew relatively fast and will be negative otherwise. Overall, the commodity composition and market distribution effects capture the fact that a country may exceed world growth rates without actually gaining market share for any particular commodity or market destination by concentrating in commodities or market destinations that are growing faster than world averages.
List of Authors
Prof. Bernadette Andreosso-O’Callaghan, Department of Economics and Euro-Asia Centre, University of Limerick, Ireland. Maria Bas (visiting scholar), Penn State University, USA. Hien Minh Dinh, CIEM-Ministry of Planning and Investment, Vietnam. Dr. Martin Falk, Austrian Institute of Economic Research WIFO, Austria. Dr. Wilhelm Kohler, University of Tübingen, Germany. PhD. Ivan Ledezma, University Panthénon-Sorbonne and Paris School of Economics, France. Svetlana Ledyaeva (PhD. Economics, Finland.
candidate),
Helsinki
School
of
Prof. Inmaculada Martínez-Zarzoso PhD., Department of Economics, Georg-August University Göttingen, Germany, and Instituto de Economía Internacional, Universidad Jaime I, Spain. Anh Duong Nguyen, CIEM-Ministry of Planning and Investment, Vietnam. Ass. Prof. Françoise Nicolas, Institut Français des Relations Internationales and Université Paris-Est, Marne la Vallée, France. Dr. Felicitas Nowak Lehmann Danzinger, Ibero-America Institute for Economic Research, Georg-August University Göttingen, Germany. PD. Dr. Özlem Onaran, Institute for Labour Economics, Vienna University of Economics and Business, Austria. Thien Hoang Pham, CIEM-Ministry of Planning and Investment, Vietnam.
264
List of Authors
Prof. Dominick Salvatore, Fordham University, USA. Thi Anh-Dao Tran, CARE - University of Rouen and CEPN University of Paris 13, France. Quang Long Trinh, CIEM-Ministry of Planning and Investment, Vietnam. Tri Thanh Vo, CIEM-Ministry of Planning and Investment, Vietnam. Sebastian Vollmer, Ibero-America Institute for Economic Research, Georg-August University, Germany. Mag. Yvonne Wolfmayr, Austrian Institute of Economic Research WIFO, Austria. Prof. Rodrigo M. Zeidan, School of Applied Social Sciences, Unigranrio, Rio de Janeiro.