India s External Linkages and Trade: An Analysis of Value Chains and the Effect of Exchange Rate

Transcription

1 India s External Linkages and Trade: An Analysis of Value Chains and the Effect of Exchange Rate By Simi Thambi Thesis Supervisor: Hiroyasu Uemura This thesis has been submitted for Doctoral Degree in Global Economics at Yokohama National University i

2 Table of Contents Acknowledgements... iv List of Figures... v List of Tables... vi Chapter 1 Introduction... 1 Chapter 2 The Importance of Vertical Production Networks and Intermediate Goods Trade in Asia: A Literature Review Introduction: Terms and Concepts Growth of Vertical Production Network Trade and Vertically Specialized Intermediate Goods Trade Reason for the Existence of Production Networks: Fragmentation Theory Ways of Fragmenting Production in Production Network Trade: Vertical Intra-Industry Trade and/or Vertical FDI Result of Fragmentation: Growth of Trade in Intermediate Goods Trade or Vertically Specialized Intermediate Goods Trade? Methodology to Quantify Vertical Specialization Foundation Measure of Vertical Specialization Value Added Trade and Vertical Specialization State of Existing Research on Value Added Trade and Vertical Specialization Discussion on Stehrer (2012, 2013) and Koopman et al. (2010, 2012 and 2014) Role of Japan and China in Vertical Production Networks and Intermediate Goods Trade Conclusion and Scope for Future Research Figures and Tables Appendix A Numerical Example Comparing Methodology of Koopman et al (2012, 2014) with Strehrer (2012, 2013) Chapter 3 India s Intermediate Goods Trade in the Inter-regional Value Chain: An Examination based on Trade Data and BRICS Input-Output Table Introduction Motivation for Research Literature Review on Intermediate Goods Trade and Value Chain Methodology and Data Trade Data Analysis Input-Output Analysis India s Intermediate Goods Trade Using Trade Statistics ii

3 6. Input-Output Analysis of India s Intermediate Goods Trade Conclusion Figures and Tables Chapter 4 External Linkages and Value Added in Imports of India: Using World-Input Output Data Introduction Data Description Methodology External Backward Linkages Value Added in Imports to India Trade in Value Added Results External Backward Linkages Value Added in Imports to India Trade in Value Added Conclusion Figures and Tables Appendix A: WIOD Industry Classification Appendix B: Linkages for all 35 Industries Chapter 5 Examining the Effect of Exchange Rate on Bilateral Trade Balance Introduction Literature Review Model and Variables Model Data Description Steps in Empirical Method Empirical Results Conclusion Figure and Tables Appendix A : Johansen Cointegeration Test Chapter 6 Conclusion Bibliography iii

4 Acknowledgements I would like to express my deepest gratitude to my academic advisor, Professor Hiroyasu Uemura, for his stimulating insights and encouragement throughout the course of my study. Without his mentorship and useful comments, this thesis could not have been completed. His kindness has always motivated me to do my best. I would also like to thank Professor Yuichi Hasebe whose classes played an instrumental role in helping me learn more about Input Output tables. I must also thank Professor Kiyotaka Sato and Mr. Nagendara Shreshta for their classes, which turned out to be very beneficial for my research work. Many thanks also to Professor Midori Kizaki, Professor Craig Parsons and Associate Professor Taku Ishiro for their insightful comments on different aspects of my research work. Thanks also to Professor Masaaki Kuboniwa (Hitotsubashi University) and Mr. Satoshi Inomata (Institute of Developing Economies, JETRO) for their beneficial comments. I am very grateful to the Japanese Government (MEXT) for funding my research as a PhD student in Japan. Finally, none of this could have been possible without the unconditional support and encouragement of my wonderful family and friends. The joy of having them in my life has made this journey so much easier. Above all, I thank God Almighty for giving me the strength and intelligence to pursue this degree and stay focussed through it all. iv

5 List of Figures Chapter 2 Figure 1 Schematic Illustration of Keywords Based on a Literature Review Figure 2 The Fragmentation Theory: Production Blocs and Service Links Figure 3 Fragmentation between the Parent and Host Country in East Asia: Vertical Intra Industry Trade and Vertical FDI Figure 4 Flowchart of Literature on Production Networks and Growth of Trade Figure 5 Koopman et al Methodology Figure 6 Import Content in Chinese Exports to the US Chapter 3 Figure 1 Global Matrix of Imported Inputs to Produce Figure 2 Global Sourcing Strategies Matrix for Domestic Production -Comparison of 1995 and 2009 Figure 3 Distribution of Employment and GDP by sector, China and India, 2000 Figure 4 Schematic Representation of (I- A) Matrix of 2005 BRICS IO Table Chapter 4 Figure 1 Disaggregated External Backward Linkages of India (3 sectors) Figure 2 External Backward Linkages of India 1995 Figure 3 External Backward Linkages of India 2011 Chapter 5 Figure 1 Trade Deficit of India ( ) Figure 2 Bilateral Exchange Rate of India with its Main Trading Partners Figure 3 Bilateral Trade Balance of India with its Main Trading Partners v

6 List of Tables Chapter 2 Table 1 VS Exports as a Share of Merchandise Exports Table 2 Different Input-Output Datasets Table 3 Simple Two Country 1 Unit Output IO Representation Table 4 Differences between VAiT (Value Added in Trade) and TIVA (Trade in Value Added) Chapter 3 Table 1 Geographic Distribution of World Trade in Goods According to Production Stage Table 2 Importance of Primary, Intermediate and Final goods in India s Trade with the World Table 3 Share of Each Industry in Total Intermediate Goods Trade of India (According to Industry Classification) Table 4 Geographical Distribution of India s Trade in Intermediate goods: Share of Each Country in India s Total Trade in Intermediate Goods Table 5 India s Intermediate Good Import from China According to Industry Classification Table 6 Power of Dispersion of Top 4 Industries: Result of Input-Output Analysis Table 7 Sensitivity of Dispersion of Top 4 Industries: Result of Input-Output Analysis Table 8 Top 4 Backward Linkages of India with BRICs+ 3 Countries: Result of Input-Output Analysis Chapter 4 Table 1 Aggregated External Backward Linkages of India (3 sectors) Table 2 Aggregated External Backward Linkages of India ( 5 sectors) Table 3 Disaggregated External Backward Linkages of India (5 sectors) Table 4 External Backward Linkage for India 1995 Table 5 External Backward Linkage for India 2011 Table 6 Foreign Value Added in Indian Imports (3 sectors) Table 7 Foreign Value Added in Indian Imports (3 sectors) Disaggregated Table 8 Foreign Value Added in Indian Imports (5 sectors) Disaggregated Table 9 Domestic and Foreign Value Added in Indian Imports from China(35 Sectors) (1995) Table 10 Foreign Value Added in Indian Imports from China (35 Sectors) Disaggregated Table 11 Domestic and Foreign Value Added in Indian Imports from China (35 Sectors) (2011) Table 12 Foreign Value Added in Indian Imports from China (35 Sectors) 2011, Disaggregated Table 13 Comparing Gross Trade Balance and Value Added Trade Balance Table 14 Value Added Trade Balance (5 sector case) vi

7 Chapter 5 Table 1 India`s Aggregate Trade Balance (Merchandise): (US $ Million) Table 2 India`s Trade Balance with major trading partners 2013 Table 3 Calculated F-Statistic for Different Lag Length Imposed on All First-Differenced Variables Table 4 Calculated t-statistic from Conditional ARDL Equation Table 5 Short Run Response of Trade Balance to Exchange Rate Table 6 Estimated Long Run Coefficients Normalized on the coefficient of TB (-1) Table 7 Short Run Response of Trade Balance to Income Variables Table 8 Comparison of Results Across Existing Literature vii

8 Chapter 1 Introduction India gained its independence from the British in From 1947 for nearly four decades until 1991, it pursued inward oriented socialist style development policy with little reliance on external trade. However, the lacklustre growth in those decades combined with a balance of payment crisis in 1991, motivated India to initiate trade liberalization reforms from The new economic reforms were geared towards deregulating the industrial sector and integrating the Indian economy with the rest of the world. Following the economic reforms of 1991, Indian economy has enjoyed steady growth over the last two decades. From the infamous Hindu growth rate of 3% in the middle of 1980s, India grew at the rate of 7% to 8% in the first decade of the 21 st Century. It now is one of the fastest growing economies in Asia. Alongside the rise of India in Asia, there have been significant changes in the way in which many Asian countries engage in trade, mainly because of the growth of global and regional value chains after the 1990s. In synch with these changes, India`s external linkages may have changed. This thesis makes an attempt to examine these changes. The objective of this study is twofold. First, we examine the changes in external linkages of India using Input- Output (IO) Analysis. In particular, we use the following methodologies: Rasmussen (1957) index to look at the external backward linkages; the VAiT (Value Added in Trade) method of Koopman (2010, 2012 and 2014) to measure value added in Indian imports disaggregated by country of origin; TIVA (Trade in Value Added) method of Johnson and Noguera (2012) to measure the value added trade balance of India and its main trading partners. Second, we examine the responsiveness of trade to exchange rate changes. We use the Auto-regressive Distributed Lag Model (Pesaran et al 2001) to analyze the short-run and long-run response of bilateral trade balance to changes in bilateral exchange rate of India and its main trading partners. The reasons why these objectives are pursued to understand the characteristics of India s external linkages and trade in this research are as follows: 1. Growing Importance of Value Chains in Asia: A value chain can be defined as the different stages in the production of a good which add value to the final product. Each of this stage can be called a part of the value chain. The term was popularized by Michael Porter (Porter, 1985). In recent decades, there has been an increase in shifting some stages of the value chain to locations abroad. This phenomenon has 1

9 been addressed in various names. slicing up the value-added chain (Krugman 1991); offshoring ; outsourcing ; fragmentation (Jones and Kierkowski 1990 and 2001, Deardorff 2001); vertical specialization (Balassa 1967 and Hummels et al. 1998, 2001, Gonzalez 2012) and unbundling (Baldwin 2006, 2012). It is surprising that there is barely any literature that examines how an emerging economy like India is linked to or benefits from these production networks. The first part of this thesis Chapter 2 to Chapter 4 begins by measuring the backward linkages of India in the global value chains for different industry classifications. It then makes a fresh attempt at quantifying the value added trade of India with its main trading partners. Quantifying value added trade is very important because it gives the true measure of the extent to which a country depends on other countries. For example, Indian imports from China could contain a high degree of foreign value added from other countries. The methodology developed in this thesis can quantify how much of third country value added is embodied in Indian imports from China. For this part of the thesis, we use Input-Output analysis as the research methodology. Use of Input-Output (IO) tables for analysis has an important advantage over the use of the trade statistics.while trade statistics only show direct interconnection among sectors, IO analysis shows direct and indirect interconnection of inputs from various sectors. Therefore, the results from IO analysis can provide a better picture of the interconnections across different industries. 2. Response of Trade to Exchange Rate Changes: The second part of this thesis deals with another important topic for India, the effect of exchange rate changes on trade balance with main trading partners. India has had a persistent trade deficit with its main trading partners in spite of a depreciating rupee. Therefore, this creates an important research agenda for analyzing the responsiveness of its trade flows to exchange rate changes. Chapter 5 examines the effect of exchange rate on India`s bilateral trade balance with its main trading partners using Auto Regressive Distributed Lag model of Pesaran et al (2001). The advantage of this method over older versions of error correction models is that ARDL model can be used even when the variables include a mixture of I(0) and I(I) variables. In the remainder of this introductory chapter, we present a synopsis of each chapter included in this thesis. 2

10 Trade in intermediate inputs has been steadily growing facilitated by fragmentation of production; increase in importance of outsourcing; and foreign direct investment. Current trade pattern in Asia is best explained by a rise in interconnected vertical network trade in intermediate goods extending across various countries of East Asia. It can be seen as a conveyor belt which passes through many countries, with each country adding value to the good at various upstream or downstream stages of production. As a result, studies on intermediate goods and value chain are attracting more and more attention. Given this background, Chapter 2, The Importance of Vertical Production Networks and Intermediate Goods Trade in Asia: A Literature Review presents a literature review of production network trade for Asian region, as a whole. It reviews the most cited literature to highlight the importance of vertical production networks in Asia. Furthermore, it reviews and links the main empirical methodologies used in existing literature to quantify the degree of vertical specialization. The next two chapters, Chapter 3 and Chapter 4, examine the changes in external linkages of India with the rise in production networks. There is a scarcity of literature that examines this issue. Few existing researches like UNENSCAP (2011), Yamashita (2012), Athukorala (2008, 2012) analyse India` s participation in production networks using trade data. However, as we mentioned before, using only trade data has its limitations. To set the stage for research, Chapter 3 India`s Intermediate Goods Trade in the Inter-regional Value Chain: An Examination Based on Trade Data and BRICS data, uses both trade data and Input-Output analysis to examine India s sales and purchases from foreign countries and brings out the advantages of Input-Output tables. For the Input-Output analysis, this chapter measures India`s external backward linkages using Institute of Developing Economies (IDE) BRICs 2005 Input- Output table. This IO table covers 25 industries for Brazil, Russia, India, China, US, Japan, EU and Rest of the World. One limitation of using BRICS table is that it is only provided for the year Therefore, we cannot use it to see the transition in linkages between two periods. The next chapter overcomes this limitation by using the World Input- Output Database. Chapter 4, External Linkages and Trade in Value Added: Using World-Input Output Data, extends the analysis of the previous chapter by using the World Input Output Table for 1995 and The WIOD tables with their wide time series coverage facilitates a more meaningful analysis. The World Input-Output Database provides the values of input-output transactions among 35 industries for 27 EU countries and 13 other major countries in the 3

11 world for the period from 1995 to The objective of this chapter is to investigate the change in external linkages of India between the period 1995 and Given the rise in trade through production networks in Asia, we want to investigate if there has also been a change in the procurement patterns of India over this period. Furthermore, using the recent methodologies of trade in value added, we make a fresh attempt to quantify the source (country of origin) and type (domestic or foreign) of value added in Indian imports. We also measure the trade balance in value added terms of India and its main trading partners. An investigation of value added trade is important because it provides useful insights into the way in which a country is linked to and benefits from the production networks. The last chapter of this thesis, Chapter 5 Examining the effect of Exchange Rate on Bilateral Trade Balance examines the effect of changes in exchange rate on trade balance of India. India`s trade balance has been in deficit for a long time, in spite of a depreciating rupee, this raises an important research agenda of how responsive are Indian trade flows to exchange rate changes? Therefore, in this chapter we examine the responsiveness of India s bilateral trade flows to changes in exchange rate. In the literature on this field, often aggregate trade balance is used to see the responsiveness of trade to exchange rate. We argue that aggregate trade balance is not a good representation to examine the effect of changes in exchange rate on trade flows. This is because aggregate trade balance does not account for the bilateral differences in exports and imports across trading partners. Thus, in this chapter we examine the ability of exchange rate to affect bilateral trade balance of India with its six main trading partners, United States, China, United Kingdom, Germany, Japan and South Korea using monthly time series for the period 1991 to To do this, we use Bonds testing method and the Auto Regressive Distributed Lag Model (ARDL) developed by Pesaran et al (2001). Finally, in Chapter 6, Conclusion, we summarize the main results of all the above chapters and interpret the main findings of this research. We also present the limitations of this research and based on that we identify the scope for future research. 4

12 Chapter 2 The Importance of Vertical Production Networks and Intermediate Goods Trade in Asia: A Literature Review Abstract The nature of trade and production in East Asia has undergone many changes over the past few decades. While a lot of researchers have examined the nature of these changes, there have been only a few attempts at reviewing their research results to determine the current state of research. The objective of this chapter is to fill this important gap by linking the methodologies and results of most cited research in this area. This review underlines the rising importance of vertical production networks and intermediate goods trade in Asia. The main methodology used in literature to quantify this vertical specialization of trade is through an analysis of international input-output tables. Such an analysis of vertical specialization through input-output tables for East Asia shows that though China takes the position as the main supplier in East Asia, there is high degree of import content in Chinese exports to the world. This chapter also identifies the scope for future research. This review finds that given the development of new and improved methodologies for quantifying vertical specialization and value added trade, there is a need to relook at many conventional approaches and indices of macroeconomics using value added exports data rather than the commonly used gross exports data. Keywords: East Asia, Fragmentation, Vertical Specialization, Value Added Trade 1. Introduction: Terms and Concepts The current trade pattern in Asia is best explained by interconnected vertical networks of trade in goods extending across various countries, with each country adding value to the good at various upstream or downstream stages of production. 1 Given this background, the aim of this chapter is to review some of the most cited literature in the field of vertical production 1 This is different from the regional transmission pattern of Flying Geese (FG) theory originally developed by Akamatsu (Akamatsu, 1962) and it extension by Kojima (Kojima, 2001) to explain the trade patterns of the early decades of East Asian industrialization. The idea of third stage of FG theory of regional transmission involved: Trade in final goods Inter- Industry trade Country as the unit of Analysis In contrast, the current trade patterns in East Asia are characterized by: Trade in intermediate goods Intra- Industry trade Firm as the unit of analysis For more on this debate, refer Uemura (2014). Uemura (2014) brings out a clear distinction between Akamatsu s original version of FG theory and its extension by Kojima. While Akamatsu s version saw trade in the Ricardian model of trade. Kojima s extension saw FG theory in a Hecksher-Ohlin framework with factor price differences and factor price equalization. He argues that the original Akamatsu s FG theory with wage differentials is more appropriate to explain the current pattern of production networks trade in Asia. 5

13 networks of trade to give the reader a concise summary of the main debates of the literature in this field, as well as identify the various methodologies used to quantify this phenomenon. There are some terms that are often used in the literature on this field. Figure 1 provides a schematic illustration of the distinction of these terms in the context of production networks. The figure has three columns: Trade, FDI and production networks in East Asia. Trade is defined as export or import of goods. Foreign Direct Investment (FDI) is defined as investment in foreign countries through licensing or setting up affiliates. The Figure 1 shows that the first set, International trade can be of two types inter-industry trade or one way trade and intra-industry trade. Inter-industry trade refers to the trade of goods among different industries of different countries. For e.g. if there are two countries, country 1 and country 2; trade of goods from the agriculture industry of country 1 for goods from the machinery industry of country 2. Another kind of trade is intra-industry trade. For e.g. goods from country 1 s machinery industry are traded for goods from country 2 s machinery industry. Intra industry trade can further be divided into horizontal and vertical trade. Horizontal trade is the case where goods are differentiated by attributes. Vertical trade is the case where goods are differentiated by quality. The second set in Figure 1 shows FDI. Some firms find it profitable to undertake foreign direct investment in other countries. They do this either by licensing to the firms in the host country or by setting up the affiliates of the parent firm in the host country. 2 In either case, the nature of FDI can be horizontal or vertical. Horizontal FDI refers to the case when firms enter the host country with a market seeking objective. Vertical FDI refers to the case when the firms enter the host country with the objective of saving costs. 3 A review of literature on production networks in East Asia shows that the definition of production networks in Asia does not fit completely into just one of these sets.i.e. only trade or only FDI. The nature of production network in East Asia includes components of both these sets. Production network trade in Asia can be seen mainly as the subset of vertical intra industry trade and vertical foreign direct investment. 4 Intra- Industry trade because 2 For details refer Ownership, Location, Internalization Theory( OLI Theory) (Dunning, 2001) 3 Some authors argue it is difficult to clearly distinguish between these two types of FDI as sometimes both these kinds of FDI coexist. For e.g. Japanese automobile assemblers are taking advantages of regional trade liberalization programs to set up production facilities in ASEAN countries and facilitate the division of labor within the region, in order to achieve a regional scale of production (UNESCAP, 2011). Nevertheless, we believe the initial motive of entry is Vertical FDI even though at some point of time later the MNC may engage in horizontal FDI. Furthermore most literature on IPN continues to vertical FDI an important feature of IPN and not horizontal FDI. 4 Vertical FDI eventually expands to include horizontal FDI in some cases. 6

14 much of the East Asian network trade deals with the parts and components of the machinery sector. The rest of this chapter is organized as follows. Section 2 presents a review on the growth of production network trade and vertically specialized intermediate goods trade. Section 3 reviews the methodology of research used in literature to quantify vertical specialization. Section 4 looks at the role of Japan and China in these vertical production networks. Section 5 concludes all the above sections and identifies scope for future research. 2. Growth of Vertical Production Network Trade and Vertically Specialized Intermediate Goods Trade Origins of international production networks have been traced to activities of MNCs from the United States. However, until the 1980s the nature of this trade was mainly a north- north trade between North American and European countries. Overtime these firms began to explore opportunities in North South trade in the neighboring countries of Latin and South America. Unfavorable investment climate in those countries induced these MNCs to look towards East Asia and South East Asia (UNESCAP; 2011). Consequently there was a rapid increase in production networks in East Asia by middle of 1990s facilitated by the favorable investment and trade policies of these countries. There are a lot of researches to investigate the new patterns of trade in East Asia. This section attempts to link the findings of some of the most cited research on fragmentation and vertical production Reason for the Existence of Production Networks: Fragmentation Theory The reason for the existence of production network trade is best explained by the theory of fragmentation of a firm. The pioneer attempt to explain this theory was made by Jones and Keirzkowski (1990). His idea is presented in a schematic illustration as shown in Figure 2. Figure 2 shows the case of a firm before and after fragmentation. The upper part shows the case before fragmentation. This figure consists of a big rectangle which is divided into smaller rectangles and squares. Assuming the big rectangles represents a firm that produces a single product. The smaller squares and rectangles are production blocs (PB). Production blocs represent the different kinds of tasks that go into producing that product. Each bloc can have different characteristics; it can be purely capital intensive, labor intensive or a mix 5 Ando and Kimura (2011) make a similar attempt at reviewing literature. However it is felt that their reviews fails to include some important researches. Even though some debates may appear the same, the value addition of this paper over the aforementioned paper is that it tries to include most cited important researches in this field. 7

15 of both. Under some circumstances, it is economical for the firm to separate these production blocs and locate them elsewhere in different host countries. The lower part of the figure shows the case of a firm after fragmentation. Each production block is now separated and linked to each other through arrows which are called service links (SL). Firms will chose to fragment their production process into production blocs if it is technically feasible and economically profitable. Technical feasibility refers to the technical separability of the production bloc from the main production process. Economic profitability means the separation of production blocks is more profitable than when in a single integrated production process framework. This happens because of two reasons. Either when there is a high efficiency of production costs due to the location advantages of the host country to which the production bloc is moved or when the service link costs incurred in connecting remotely located production blocs are small. These service link costs depend not only on trade barrier and transport cost but also on various coordination costs. 2.2 Ways of Fragmenting Production in Production Network Trade: Vertical Intra- Industry Trade and/or Vertical FDI In the context of production networks, there are two main ways in which the firm can fragment its production blocs. The literature on this field points towards two main activities, vertical intra industry trade and vertical FDI. Figure 3 explains this idea using the example of 2 countries, parent country and host country 6. Parent country can fragment its production to the host country in two ways via arms length transactions with host country firms or by setting up multinational affiliates of the parent firm in the host country. The choice of internalization is primarily the parent firm s choice 7. Figure 2 and Figure 3 taken together explain in a simple framework, the interactions between 2 countries. In reality however, the idea is much more complicated with many countries and intricate networks of firms. 6 Ando and Kimura (2011) in Figure 4 provide an illustration of two dimensional fragmentation 7 For details on this refer literature on the theory of internalization of OLI paradigm(dunning, 2001) 8

16 2.3 Result of Fragmentation: Growth of Trade in Intermediate Goods Trade or Vertically Specialized Intermediate Goods Trade? Fragmentation involves the breakup of the production process into fragments of tasks that produce different kinds of intermediate goods 8 which when joined together result in the production of the final good. Many intermediate goods can go into the making of a final good. Moreover, each intermediate good can further be used in the production of other intermediate goods. This creates an interconnected network of production extending across different countries. Vertically specialized intermediate goods can be defined as a subset of intermediate goods, those goods which cross multiple countries for different stages of production. 9 Trade in intermediate goods has increased significantly in the past two decades. Some critics argue that a large part of the growth in world trade in the past decades can be attributed to the growth in the trade of intermediate goods. Trade flows are dominated not by goods that are fully consumed but by goods that are further used in the production of other goods and services. Intermediate inputs represent 56% of total goods trade and 73% of total services trade in The trade of intermediate inputs has grown at an average annual rate of 6.2% for goods and 7 % for services (in volume terms) between 1995 and However, trade in final goods and services has increased at the same pace and as a consequence, the share of intermediate goods trade in total trade has remained constant while the share of intermediate services trade has slightly increased (Miroudot et al. 2009). Other studies also confirm the growing importance of intermediate goods trade (Ng and Yeats 2001, 2003; Athukorala and Yamashita, 2006). However, some critics argue that main reason for growth of world trade in the past decades is the increase in the trade of a subset of intermediate goods that they call vertically specialized intermediate goods. They argue that the share of intermediate goods in total trade has remained constant. In fact, it is the share of vertically specialized intermediate goods, those intermediate goods that cross multiple borders that has been increasing. 8 An intermediate good is defined as an input to the production process that has itself been produced and, unlike capital, is used up in production. The difference between intermediate and capital goods lies in the latter entering as a fixed asset in the production process. Like any primary factor (such as labor, land, or natural resources), capital is used but not used up in the production process. On the contrary, an intermediate good is used, often transformed, and incorporated in the final output. As an input, an intermediate good has itself been produced and is hence defined in contrast to a primary input. As an output, an intermediate good is used to produce other goods (or services) contrary to a final good which is consumed and can be referred to as consumption good ). For e.g. the production of a car require tires which requires steel and rubber; steel requires iron which requires iron ore and so and on, similarly rubber requires other inputs. This is a definition provided by Deardorff (2006) in his Glossary of International Economics 9 A good is called vertically specialized when three conditions are met: (1) a good (or service) is produced in two or more sequential stages; (2) two or more countries provide value-added during the production process; and (3) at least one country uses imported inputs in the process and some of the output is exported (OECD and WTO, 2012). 9

17 Figure 4 shows the flow chart of the scope of literature on production networks in the order of their importance in production network trade. The literature on production networks flows from intermediate goods to vertical specialized intermediate goods. In other words, within the trade in production networks, the most important category that has led to growth of trade is intermediate goods and even within that it is vertically specialized goods that are most prominent. Yi (2003) in his famous magnification effect analysis asserts that it is vertically specialized goods which cross multiple borders that explain most of the growth of world trade. Magnification effect is the magnifying effect of tariff reductions into large increases in trade when goods cross the borders of multiple countries. For instance, if there are N sequential stages each in a different country, a one percent decrease in tariff will cause the cost of production to reduce by N% (in contrast to a 1% decline in the cost of a regular traded good). This is the reason why the degree of increase in world trade over the past few decades is much more than can be explained by decrease in tariffs, in other words the extent of increase in world trade is much larger than can be explained by the tariff reduction in standard trade models. He shows this by including the idea of vertical specialization in a 2 country dynamic Ricardian trade model with three stages of production: primary, intermediate and final goods. Bridgman (2012) agrees with the above and extends his study to argue that magnification effect is most significant in manufactured intermediate goods. In his study, he presents a three stage vertical specialization Ricardian trade model of raw materials, manufactured parts and final goods sectors by looking at the exchanges between the home and foreign country with two households in a tractable general equilibrium framework for world trade, during the period Through his simulated trade model he looks at growth in world trade and changes in the composition of trade. At the sector level, manufacturing trade tripled while overall trade growth only doubled. Within the manufacturing sector, in the category of intermediate goods, his results show that raw materials share of intermediate goods trade have decreased while the share of manufactured goods in intermediate goods has increased. Within the share of manufactured goods in intermediate goods, the share of vertically specialized goods is rising rapidly. Even though at the overall level, the share of intermediate goods in total trade has been constant. He avers that lower trade costs on manufactured parts led to VS trade growth. Chen et al (2005) also find that trade of intermediate goods to total trade has not increased but a subset of trade in intermediate goods which deals with the trade of vertically 10

18 specialized goods has increased. Table 1 shows vertically specialized (VS) exports as a share of merchandise exports has increased for 10 OECD countries for the period 1968 to According to another prominent study by Yeats (2001), vertically specialized intermediate goods trade is most prominent in the parts and component industry. Yeats (2001) in his pioneer study looks at the commodity composition of global production sharing. His results pointed towards the importance of trade in components within the entire machinery and transport (SITC 7) sector. To sum up, in this section, we saw that the reason for production network trade is best explained by the fragmentation theory of production that splits production blocs across national borders. The firms mainly use vertical FDI and vertical intra-industry trade to fragment production. Finally, we highlighted the growing importance of vertical specialization in intermediate goods of parts and components facilitated by reduction in tariffs and transport costs. In the next section, we will review the main methodology used in literature to quantify this phenomenon of vertical specialization. 3. Methodology to Quantify Vertical Specialization With the rising importance of vertical specialization in international trade many researchers have identified ways to quantify vertical specialization through an analysis of international input- output tables. Table 2 lists the main databases used by researchers in this field. The development of literature on quantifying vertical specialization originated from the pioneer work of Hummels et al (2001). In the years that followed, this method was extended by several researches like Johnson and Noguera et al. (2012, 2014), Koopman et al (2012, 2014) and Strehrer (2012, 2013). The objective of this section is to review the main results and link the methodologies used by all the above mentioned authors. 3.1 Foundation Measure of Vertical Specialization Hummels et al (2001) made one of the first attempts to quantify vertical specialization. There are two ways in which a country can be engaged in vertical specialization - By using imported intermediates to produce exports By exporting intermediate goods which are used as inputs by other countries to produce goods for export. 11

19 He represented the former as VS and the later as VS1. However, he developed a clear mathematical formula only for the former leaving the latter for future research. 10 His model was formulated as a two country case for vertical specialization. Such a model can be used to look at the interaction of two countries with each other or one country interaction with the world aggregated as the other country. For two countries r and s and one sector, vertical specialization (VS) in terms of import content of export of country r can be formulated as: VS r= ( imported intermediates sr ). exports gross output rs r This equation shows the ratio of imported intermediates in the total output of country r multiplied by the exports from country r to s. For the case of n sectors, aggregate Vertical Specialization (VS) for country r is the sum of imported intermediates from country s to r across all sectors n divided by total exports of all sectors for country r, Aggregate VS r = n VS n n Xr = n VS r n X r = μam X X n where, X n is a scalar representing the sum of country r exports And VS r = ( imported intermediates sr n n ). exports gross output r = μa M X r Where, μ is a 1 x n unit vector. A M is the n x n imported coefficient matrix X is an n x 1 vector of exports, n is the number of sectors In the above equation μam X X n captures the direct import content of exports (expressed as the share of intermediate imports over total exports) but it does not reflect how imported intermediates are used indirectly throughout the economy. To capture the indirect use of imported intermediates at different stages of domestic production, before these are embodied 10 VS1 was clearly formulated by his successor Koopman et al. (2010, 2012 and 2014) 12

20 in an export flow, Hummels introduced the Leontief inverse matrix [I - A D ] -1 to the above expression. Here A D is an n x n domestic technical coefficient matrix which reflects the value of domestic intermediates over total output (rather than imported intermediates over imported intermediates which was used in the case in the A M matrix). Adding this Leontief inverse to the above equation captures the use of imported intermediates across each stage of domestic production allowing the imported input to be embodied in a domestic output at the 2 nd, 3 rd, 4 th,... stage before it becomes embodied in the good that it is exported. Therefore, finally the HYI VS Index (Hummels et al. (2001) Vertical Specialization Index) can be written as: HYI VS Share of Total Exports = μam ( I A D ) 1 X X r (1) where, u is a 1 x n unit vector, A M is the n x n imported coefficient matrix, I is the identity matrix, A D is the n x n domestic coefficient matrix, X is an n x 1 export vector, X r is the total country exports and n is the number of sectors. The above mentioned HYI VS index made the hallmark attempt at quantifying the import content of export or VS, a sequence of vertical specialization which was hitherto unmeasured. However, it had some limitations which are as follows: HYI VS captures the import content of export assuming only one country exports intermediate goods. Therefore, whenever both the country s export intermediate goods, the model diverges from the real import content of export. It uses the assumption that imports are 100% foreign sourced, in other words, it does not take into account that some inputs might be reflected back after processing abroad. HYI VS measure is an aggregate measure that shows the intermediate imports from all countries into home country embodied in the exports to all countries. However, it does not measure the bilateral degree of vertical specialization. The researches that followed his work, improved upon the limitations in the following ways: To overcome the first limitation, the definition of exports was broadened to include not only export of final goods (as in the case of HYI VS measure) but also the export of intermediate goods by both countries. 13

21 The second limitation arose because in the case of HYI VS, only the import coefficient matrix was used to calculate the direct import content of export, assuming all imported content was foreign value added with no domestic value added and all the foreign value originated in one country. However, this is not the case in the real world. We can show this with an example. Consider the following three cases: Case 1: Direct Import - For instance, US imports from China a good that is completely made in China, 100% Chinese value added. Case 2: Import includes goods reflected back to home country- For instance, US imports from China a good that has Chinese value added and US value added, that returns back to US after being processed in China. Case 3: Import includes reflected back imports plus indirect imports: In the Case 2 above, if Chinese value added consists of parts imported from other Asian countries, then that means US is indirectly importing the value added of those countries. To measure the above mentioned sequences of trade, researches identified that it is best to use the direct domestic value added vector. This is one of the main differences between older literature on vertical specialization and newer literature on value added to measure vertical specialization. To overcome the third limitation, i.e. to change the HYI measure from aggregate to bilateral, two methods have been used in literature. One way is by using the proportionality assumption. The proportionality assumption assumes that the intensity in the use of intermediate inputs is the same between production for exports and production for domestic sales. The other way is to use detailed end use classification (intermediate or final) of detailed import statistics like BEC (Business Economic Classification). The second method is increasingly used by researches rather than the conventional proportionality assumption. 3.2 Value Added Trade and Vertical Specialization State of Existing Research on Value Added Trade and Vertical Specialization Given the limitation of the HYI VS measure mentioned in the previous section, the researches that followed identified that the best way to look at vertical specialization was to 14

22 look at it through value added in trade, i.e. by making use of the value added coefficient vector provided in any input-output table 11. The following equation explains the link: Table 3 attempts to link direct value added vector and import coefficient vector used by HYI VS using a two country simple Input-Output Model: In a country with just 2 sectors u is a 1 2 row vector, u=(1,1) [ A 11 A 12 ] is the input coefficient matrix A 21 A 22 The economic explanation of this is that to produce one unit of output in country 1, country 1 requires A 11 units of domestic goods and A 21 units of imported intermediate goods. Therefore, the fraction of domestic output that represents domestic value added in Country 1 is V 1 = u - ua 11 -u A 21. Similarly, we can arrive at direct value added vector for country 2. It is easy to see from the above equation the connection between domestic and import coefficient matrix and direct value added coefficient vector. In the remaining part of this section, we will review the prominent literature that uses this direct value added vector to quantify domestic and foreign value added. Johnson (2012) is famous for his measure called VAX ratio (Value added to exports ratio). VAX ratio is defined as the ratio of direct domestic value added to gross exports. For a two country case (1, 2) with many sectors using the terminology used in his paper, VAX can be written as: 12 VAX = va 12 x 12 = r 1 [s] y 12 [s] x 12 [s] where, r 1 [s] is payment to domestic factors or domestic value added. It is the direct value added row vector. (2) y 12 [s] is called implicit output transfer from country 1 to country 2. The output from country 1 directly and indirectly used to produce final goods absorbed in country 2. It can be written as y 12 [s] = (I-A) -1 Y12, where Y12 is direct export by country 1 in the 11 A detailed examination of all the proofs of the mentioned in the literature on this field is beyond the scope of this paper. This paper will only look at the main measurement equations of different methodologies. 12 Based on the explanation of this equation, it is easy to see that VAX ratio= V(I A) 1 Y or VBY (using Koopman,2012) terminology. E= AX+ E E Y( Intermediate good export plus final good export) 15

23 form of final demand of country 2 and (I-A) -1 is the Leontief inverse which shows the intermediate inputs that were indirectly used in making Y12. x 12 [s] is the gross bilateral export from country 1 to country 2 including export of intermediate goods and export of final goods. Koopman et al. (2010, 2012, and 2014) is one of the latest methodologies in the literature on vertical specialization and value added. The value addition of Koopman et al. (2010, 2012, and 2014) to previous work is that they decomposed value added into different sequences, for the first time in literature. The literature that uses this methodology has two related branches. VAiT (Value added in Trade) and TIVA (Trade in Value Added). Before we proceed, it is important to have a clear idea of the differences between these two concepts which look very similar. Table 4 presents a concise summary of the main differences between the two approaches. For this chapter, we are mainly interested in the source of value added from the viewpoint of vertical production networks. Therefore, we focus on reviewing the Koopman (2010, 2012 and 2014) methods to quantify VAiT or Value Added in Trade. Koopman et al. (2010) contributed to the existing literature by decomposing the gross exports of a country into value added components by source. Assuming all gross output produced by country r must be used as an intermediate good or final good at home or abroad. For a two country case, gross output equation for country r can be written as follows. X r = A rr X r + A rs X s + Y rr + Y rs (r, s=1,2) where X r is N 1 gross output vector of country r, A rr is the N N domestic input coefficient vector, A rs is N N input coefficient matrix giving intermediate use in s of goods produced in country r, Y rr is the N 1 domestic final demand vector, Y rs is N 1 final demand vector in country s for final goods from country r. The two country production and trade system can be written in block matrix form as follows: ( X 1 X 2 ) = ( A 11 A 12 A 21 A 22 ) ( X 1 X 2 ) + ( Y 11 + Y 12 Y 21 + Y 22 ) 16

24 ( X 1 ) = [ I A 1 11 A 12 ] ( Y 11 + Y 12 ) = ( B 11 B 12 ) ( Y 1 ) X 2 A 21 I A 22 Y 21 + Y 22 B 21 B 22 Y 2 This system can also be succinctly expressed as follows: X = (I A) -1 Y = BY, Where B stands for Leontief Inverse Matrix (I A) -1 Using this share to create VAS (Value added share) matrix defined as: VAS=VB= ( V 1B 11 V 1 B 12 V 2 B 21 V 2 B 22 ) Where V= ( V V 2 ) is the direct value added coefficient diagonal matrix Based on this VB foundation, Koopman et al. (2012) named two measures: VAiT & TIVA. In this paper, we focus on Value Added in Trade VAiT = VBE = ( V 1B 11 V 1 B 12 V 2 B 21 V 2 B 22 ) ( E E 2 ) = ( V 1B 11 E 1 V 1 B 12 E 2 V 2 B 21 E 1 V 2 B 22 E 2 ) (3) where E rs = s (A rs final goods from country r to country s X s + Y rs ), in other words, export of intermediate goods and From equation (3), we can see two sequences, the diagonal terms V 1 B 11 E 1, V 2 B 22 E 2 are domestic value added and off-diagonal terms V 2 B 21 E 1, V 1 B 12 E 2 are foreign value added for country 1 and country 2 respectively. Similarly, in the case of three countries we can write equation (3) as follows, V 1 B 11 V 1 B 12 V 1 B 13 VB = ( V 2 B 21 V 3 B 31 V 2 B 22 V 3 B 31 V 2 B 23 ) V 3 B 33 (3a) E E = ( 0 E 2 0 ) 0 0 E 3 (3b) 17

25 V 1 B 11 E 1 V 1 B 12 E 2 V 1 B 13 E 3 VBE = ( V 2 B 21 E 1 V 3 B 31 E 1 V 2 B 22 E 2 V 3 B 31 E 2 V 2 B 23 E 3 ) V 3 B 33 E 3 (3c) where E 1 is the sum of country 1`s exports to country 2 and country 3. E 2 is the sum of country 2`s exports to country 1 and country 3. E 3 is the sum of country 3`s exports to country 1 and country 2. From equation (3c), we can see three sequences, domestic value added in exports, foreign value added in exports and indirect value added exports. The diagonal terms measure domestic value added exports (DV) DV r = V r B rr E r The sum of off-diagonal elements along a column is the true measure of value added from foreign sources (FV) embodied in a particular country`s gross exports. FV r = s r V s B sr E r The sum of off- diagonal elements along a row provides indirect value added (IV) which is a country`s value added embodied as intermediate inputs in third countries` gross exports or indirect value added exports. IV r = s t V r B rs E st The sum of a column is the sum of domestic value added and foreign value added, which is nothing but the total exports of a particular country. Its general form can be expressed as E r = DV r + FV r The above equation can further be decomposed into the following form: E r = V r B rr s r Y rs + V r B rr s r A rs X ss +V r B rr s r A rs X st +V r B rr A rs X sr s r +FV (4) 18

26 The first four terms are the components of DV r which can be understood as follows: 13 where 1. V r B rr s r Y rs 2. V r B rr s r A rs X ss ( Direct value added export) 3. V r B rr s r A rs X st (Indirect value added exports) 4. V r B rr A rs X sr = s r =Exported in final goods ( Direct value added export) =Exported in intermediates absorbed by direct importers =Exported in intermediates re-exported to third countries Exported in intermediates that return back Summing up 1, 2, 3 divided by gross exports gives the VAX ratio which is formalized by Johnson and Noguera (2012), as seen in equation (2) Summing 3 and 4 gives HYI VS1 which is mentioned in Hummels et al. (2001) Furthermore, FV r in the equation (4) is HYI VS measure Koopman et al. (2012, 2014) extended on the analysis of Koopman et al. (2010) in mainly two ways. First, they clearly identified which parts of official trade data are double counted and what is the source of this double counting. Second, they further decomposed foreign value added into its components. Figure 5 presents the extensions of Koopman et al. (2014) over Koopman et al. (2010). 14 Stehrer (2012, 2013) calculated VAiT in a different way from the Koopman et al (2010, 2012, and 2014) method. Their measure emphasizes on the net value added in trade. The main differences are summarized as follows: Firstly, the export matrix is different in Stehrer (2012, 2013) compared to Koopman et al. (2010, 2014). This can be clearly seen by looking at the export matrix for a three country case. The vector of gross trade for country 1 can be written as, E 12 + E 13 E 1 = ( E 21 ) E For details on calculation see Koopman et al. (2012) Pg.14. A detailed exposition of all the formulas is beyond the scope of this paper. 14 For details see Koopman et al. (2014) Pg. 9 to Pg

27 where, E includes both intermediate exports and final good exports denoted as E=AX+Y. Using the same logic for the other two countries, three country matrix E * can be written as follows: E 12 + E 13 E 12 E 13 Export matrix E = ( E 21 E 31 E 21 + E 23 E 32 E 23 ) E 31 + E 32 (5) Note the difference between this new E * matrix and in the case of Koopman (2010, 2014) in equation (3b). Post multiplying this new E * matrix with VBS matrix in the equation (3a), we get the Strehrer version of VB E * matrix. The second difference between the two methodologies is that the sum of the each column of E * matrix in the case of Stehrer (2012, 2013) is equal to the net trade balance. This can be easily seen from equation (5). And the sum of each column of VBE * is equal to the net value added trade balance. On the other hand, the sum of the column elements of Koopman et al (2010, 2014) VBE matrix in equation (3c) is gross exports. Thirdly, in Stehrer (2012, 2013) the sum of column elements of export matrix is equal to column elements of value added matrix. In other words, the net trade balance is equal to the value added trade balance for all countries Discussion on Stehrer (2012, 2013) and Koopman et al. (2010, 2012 and 2014) Kuboniwa (2014), using Strehrer (2012, 2013), produced an interesting result that showed the bilateral equivalence between TIVA and VAiT in net value added trade. He argued that this condition of bilateral equivalence cannot be proved using the Koopman et al (2010, 2012, and 2014) method. He thus avers that for the calculation of net value added trade, Strehrer (2012, 2013) has an advantage over Koopman et al (2010, 2012, and 2014) method. Kuboniwa (2014) result is unique in its finding; nevertheless it is important to highlight the importance of the original Koopman et al (2010, 2012, and 2014) methodology over previous methodologies. Koopman et al (2010, 2012, and 2014) method has a number of advantages over Strehrer (2012, 2013). It has the ability to quantify measures like: domestic 15 This can be seen by the expression, using the definition of value added. r t VaiT,Net = V(I A) 1 E = i` (I A)(I A) 1 E = i`e r = t Gross,Net Here i` denotes a 1 n summation vector. Vis the direct value added coefficient diagonal matrix. E* is equation(4) 20

28 value added, foreign value added, indirect value added, exports reflected back and also double counted term. Furthermore, for the first time in the literature of international trade, the above mentioned components of gross exports could be decomposed at both bilateral and multilateral level. Appendix 1 presents a numerical example to compare Koopman et al (2010, 2012, and 2014) with Strehrer (2012, 2013) method. According to OECD and WTO (2011), Koopman et al. (2011) provide a full decomposition of value-added exports in a single conceptual framework that encompasses all the previous measures. Exports are first decomposed into domestic value-added, returned domestic value added (domestic value added that comes back incorporated in foreign inputs produced with domestic inputs) and foreign value-added. Domestic value-added is then split between exports absorbed by direct importers and indirect exports sent to third countries. Between the pioneering work of Hummels et al. (2001) and these latest studies, the conceptual framework has been enhanced and we now have a full understanding of what constitutes trade in value-added terms. The exercise consists in distinguishing domestic and foreign value-added. But, as previously pointed out, the difficulty is that domestic value-added can be found indirectly in imports of foreign inputs (as returned domestic VA ) and when exported to another country can be also indirectly found in exports from third-countries. The field is therefore not only extremely relevant, but also fully mature for its full inclusion in official statistics. The next step in the analysis is to provide a full decomposition of the foreign value-added according to the country of origin of the VA. 16 (OECD and WTO, 2011; Pg20) To conclude in this section, we discussed four main methodologies that have been used in literature to quantify vertical specialization and value added trade. Hummels et al (2001) made the pioneer attempt at quantifying vertical specialization but his index included restrictive assumptions which resulted in double counting and were not applicable in the real world. His research was extended by researches like Johnson (2012), Koopman et al. (2010, 2012 and 2014) and Stehrer (2012, 2013). While Johnson (2012) framework focused only on final demand exports (F or Y). Koopman et al. (2010, 2012 and 2014) and Stehrer (2012, 2013) focused on final demand as well as intermediate exports (E=AX+F). Out of all these, 16 Chapter 4 of this thesis extends on this methodology. It distinguishes the domestic and foreign value added in Indian imports. Furthermore, it calculates third country value added in Indian imports from a particular foreign country. For more on this, please see Pg.84 of this thesis. 21

29 the most elaborate methodology seems to be that of Koopman et al. (2010, 2012, and 2014). This is because for the first time in literature the Koopman methodology was able to quantify measures like: domestic value added, foreign value added, indirect value added, exports reflected back and also double counted term. 4. Role of Japan and China in Vertical Production Networks and Intermediate Goods Trade Both Japan and China have an indispensable role in East Asian production networks. This section reviews the results of literature using the methodology mentioned in the previous section. Our objective for this section is to identify the role of Japan and China in the vertical production networks. Dean et al. (2008) quantified the import content of Chinese processing exports and normal exports using Hummel et al (2001) methodology for the period 1997 and 2002, by sector and trading partners using detailed Chinese Input Output Database for 1997 and Their main result was that China s trade has become more vertically specialized, between 1997 and 2002 China s VS share increased by about 23%. The most vertically specialized sectors in China s trade are plastics, steel processing, communications equipment, industrial machinery, metal products, and electronic computers. In these sectors, imported intermediate goods account for between 52 percent and 76 percent of the value of Chinese exports. Within that the wholly-owned foreign firms and joint ventures have the most vertically specialized trade. There is a strong evidence of Asian network of suppliers to China, Japan and the Four Tigers account for more than half of the value of China's imported inputs both in 1997 and in They identify Hong Kong, the US, and Singapore as being the most vertically specialized among China's export destinations. Baldwin and Gonzalez (2013) develop the Hummel et al (2001) to decompose the sourcing patterns of intermediate imports used by China for US exports. Figure 6 presents their result. The height of each bar shows the share of China s exports to the US made up of imported intermediates. The bars also show where the imported intermediates are sourced from. Their finding is in tandem with Dean et al. (2008). Japan, Taipei and Korea stand out as main countries from which China imports intermediate inputs for export to US This results is in synch with the theoretical argument of Yun (2000) in his study about Asian production patterns where he argued that what can be seen in East Asia is a system of hierarchical complex production links which are connected vertically backwards to Japan due to dependence on exports of Japanese technology and vertically forward to US due to its continuing position as the main external export market for East Asian manufacturers. 22

30 Johnson et al. (2014) look at the sector level export shares for China in gross and value added terms using VAX ration of Johnson et al. (2012). They find that in the Electrical and Optical equipment industry, the real value added by China is less than half of the total gross exports in that industry. Furthermore, they look at the share of Electrical and Optical Equipment in value-added and gross exports over time. The share of this sector in gross exports has almost doubled since 1995, while the value-added share has barely changed. Koopman et al. (2008, 2010, 2012 and 2014) using Dean et al. (2007) methodology of breaking down China`s total exports into processing exports and normal exports, found that domestic value added in China`s processing exports is much lower than that in China`s normal exports. For e.g. in 2004 domestic value added in China`s normal exports was 44.2% while for China`s processing exports the figure was only 28.8%. 18 Furthermore, it is interesting to see that the returned domestic value added is only 0.3% and 1.2 % for China`s normal exports and processing exports respectively. On the other hand for advanced countries like Japan, Western EU and US, this returned domestic value added share of gross exports is much greater at 2.9%, 7.4% and 12.4% respectively. In addition, Koopman et al. (2010) computed a country sector-level index for position in global value chain called GVC_positon. It is defined as log ratio of a country- sector`s supply of intermediates used in other countries` exports to the use of imported intermediates in its own production.. If the country lies upstream in a supply chain in other words is the supplier of intermediate inputs, the numerator is large. On the other hand, if it is downstream then the denominator is large. 19 Their results showed that for electronic equipment sector, Japan is one of the most upstream countries along with Western EU, US and Mexico. China on the other hand is one of the most downstream countries, indicating that its exports mostly use imported components for production. Strehrer (2013) finds that in value added terms China`s trade surplus with the US is considerably reduced from to bn in US$. On the other hand, in value added terms Japan trade surplus with US is increased from 50.7 to 67 bn US$. To conclude this section, from the above literature we can understand the importance of vertical production network trade in East Asia where the inputs for downstream stages of production network are provided by mainly upstream countries Japan and the four dragons, which are processed in countries like China for export to other destinations. 18 For detailed decomposition of gross exports of all countries for the year 2004, see Pg 32, Koopman et al. (2008, 2010, 2012 and 2014). 19 GVC_positon= Ln (1 + IV ir E ir ) Ln( 1 + FV ir E ir ) Koopman et al.( 2010) Pg

31 5. Conclusion and Scope for Future Research This chapter did a selective review on the importance of vertical production networks and the methodology used to quantify the degree of vertical specialization. In Section 2, we surveyed literature to show an increase in the degree of vertical production due to rise in production network trade. In Section 3, we reviewed the main methodologies used to quantify vertical specialization and value added trade. Specifically, we linked the results of Hummels et al. (2001), Johnson (2012), Koopman (2010, 2012, and 2014) and Stehrer (2012). In Section 4, we examined the role of Japan and China in the production networks of Asia. Finally in this section, having surveyed the existing literature in the previous sections, we highlight the scope for further research as identified in the literature reviewed in this chapter: The ability to quantify domestic and foreign value added has considerable macroeconomic implication. Therefore, it is now important to reinvestigate the conventional indices/measures in terms of value added rather than gross terms. For instance: Bilateral Trade Balances: Bilateral trade balances are generally not equal in gross and value added terms. With the increase in fragmentation through vertical production networks, use of gross trade measures like gross exports are can be misleading. This is because of substantial double counting in gross exports data which overstates the amount of domestic value added in exports. 20 Difference in composition of gross and value added export: The composition of exports in gross and value added terms may not be the same. Countries that look like dominant exporters in particular sectors may in fact contribute very little in terms of value added export. For example, India`s business services sector has a revealed comparative advantage in gross terms. However, in value added terms, it is revealed comparative disadvantage for India. (Koopman 2010, 2012, and 2014). 21 Value added exports and factor content of trade: The conventional approach overstates factor trade because it assumes that gross exports of a country are produced using that country`s technology. But with fragmentation, gross exports are produced using a combination of both foreign and domestic inputs. 20 In Chapter 4 we extend on this analysis for India. 21 Our results in Chapter 4(Table 13 and Table 14) support this. 24

32 Accounting for trade in intermediate goods lowers the traditional measure of factor content of trade that takes into account only trade in final goods. Responsiveness to exchange rate movements: The conventional measures look only at the export responsiveness to exchange rate movements. However there is a possibility that we may get different results if we use value added exports. For example, if Renminbi appreciates against all countries, how much this appreciation raises the relative price of Chinese exports depends on how much Chinese value added is embodied in its exports. One way to address this issue, will be to calculate `real effective exchange rates for value added`. With the growth of production networks, trade in intermediate goods has significantly increased. The effect of exchange rate can be different for intermediate and final goods. Therefore, the differential effect of exchange rate on trade balance of different types of commodities can be an interesting topic for future research 22. Value added can further be decomposed into its components like capital income and labor income. For instance, Timmer et al. (2013) proposed a new measure `GVC jobs` for deeper analysis of trade among European economies. `GVC jobs` is further divided into high, medium and low skilled category using International Standard Classification of Education (ISCED). This is then calculated separately for domestic and foreign value added. Timmer et al. (2014) further extended the coverage of this analysis using World Input Output Database for 40 countries. This work can be a great starting point for a similar analysis of the Asian economies including ASEAN economies, an area that has not been explored yet. To conclude, this chapter contributed to the existing literature by providing a summary of the concept of vertical production networks by reviewing the results and methodologies of the most cited literature related to vertical production network. Based on that, the main findings of the chapter are as follows: 1. A firm`s motivation for being engaged in production network is best explained by the fragmentation theory. This theory states that firms will choose to fragment their 22 Results of Chapter 5 of this thesis showed that effect of exchange on trade is in not significant on Indian trade with three out of six countries in the long run and with five out of six countries in the short run. For instance, the effect of bilateral real exchange rate was not statistically significant on India s trade balance with China in the short run as well as in the long run. India trades with China mostly in intermediate goods; this could be one reason why the effect of real exchange rate is not statistically significant in determining trade flows. This area certainly warrants further empirical examination. 25

33 production process into production blocs if the service link costs of connecting the different production blocs are economically feasible. 2. Within the trade in production networks, the most important category that has led to growth of trade is intermediate goods and even within that it is vertically specialized goods that are most prominent. 3. Out of the existing methodologies to quantify vertical specialization, Koopman et al (2010, 2012, and 2014) has the maximum coverage as it provides a full decomposition of a country`s value-added exports into domestic value added, foreign value added, indirect value added, exports reflected back and also double counted term. 4. Use of the above methodology on international input- output tables underlines the importance of China as a downstream country where intermediate inputs from upstream countries like Japan and the four dragons are processed for export to destinations like US and EU. In Chapter 4 of this thesis, we examine whether the same set of upstream and downstream countries can be identified for the case of India. 26

34

35 Figure 3 Fragmentation between the Parent and Host Country in East Asia: Vertical Intra Industry Trade and Vertical FDI Vertical FDI into the host country HOST COUNTRY PARENT COUNTRY Parent Firm MNE of Parent Firm ( Cross border intra firm fragmentation) Export/ Import Vertical Intra Industry Trade Host Country Firm (Cross border arm s length fragmentation) Source: Author Figure 4 Flowchart of Literature on Production Networks and Growth of Trade Production Networks Intermediate Goods Vertically Specialized goods Source: Author 28

36 Figure 5 Koopman et al Methodology 1. Koopman et al (2010) 2. Koopman et al (2014) 29

37 Figure 6 Import Content in Chinese Exports to the US Source: Baldwin and Gonzalez (2013) Figure 34. Using OECD IO table 30

38 Table 1 VS Exports as a Share of Merchandise Exports Source: Chen et al. (2005) Pg 42 VS exports as a share of merchandise exports 31

39 Table 2 Different Input-Output Datasets Name of Dataset Global Trade Analysis Project Database Key Features Input-output tables for over 100 countries for various benchmark years, mostly after World Input-Output Database Global tables covering OECD countries and major emerging markets from IDE JETRO Asian Input Output Regional tables covering 8 East Asian countries at five-year intervals between Tables 1985 and OECD Input-Output Tables Input-output tables for OECD countries and major emerging markets, available various years from YNU Global Input-Output Table (YNU GIO) 29 endogenous countries (out of which 11 are major Asian economies: Japan, China, Korea, Taiwan, Singapore, Malaysia, Indonesia, the Philippines, Thailand, Vietnam and India) and 59 exogenous countries with 35 production sectors. It is available for every year from 1997 to Source: Adapted from Johnson (2014) Table 1, Pg. 123 Table 3 Simple Two Country 1 Unit Output IO Representation Country 1 Country 2 Country 1 A 11 A 12 Country 2 A 21 A 22 Value Added V 1 V 2 u u Source: Author 32

40 Table 4 Differences between VaiT (Value Added in Trade) and TIVA ( Trade in Value Added VaiT (Value added in Trade) TIVA (Trade in Value Added). Definition Value Added in Trade (VaiT) measures the domestic and foreign value added content of trade. The emphasis here is on the source of value added Formula VBE where, V= Value Added Coefficient Matrix B= Leontief Inverse Matrix E= AX+ F( AX is intermediate good export and F is final good exports) For 2 countries, exports of country 1to country 2: Trade in Value added (TIVA) measures how much of value added of a particular country is contained in the consumption of another country. The emphasis here is on the destination of where value added by a particular country is consumed VBY where, V= Value Added Coefficient Matrix B= Leontief Inverse Matrix F= Final Demand For 2 countries, exports of country 1 to country 2: f 12 e 12= a 12 x 2 + f 12 Treatment of Intermediate Exports and Final good Export Includes intermediates in Exports Excludes intermediates, Includes only Final Demand. Diagonal Elements Diagonal elements show domestic Diagonal elements show value Value Added Content of Exports Added Absorbed at Home Off diagonal Elements Off diagonal elements show foreign Value Added Content of Exports Off diagonal elements show export of Domestic Value Added to other countries Column Sum Gross Exports Final Demand Compiled Databases UNCTAD- EORA GVC Database For 187 countries from WTO-OECD TIVA Database For 57 countries in 1995, 2000, 2005, 2008 and Source: Author 33

41 Appendix A Numerical Example Comparing Methodology of Koopman et al (2012, 2014) with Strehrer (2012, 2013) Intermediate Demand Final Demand Country 1 Country 2 Country 3 Country 1 Country 2 Country 3 Total Final Demand Output Country Country Country VA Output A I I-A B E=AX+F E^ V V^

42 1) Koopman et al ( 2012,2014) Methodology V^B DVA IV V^BE FVA Gross Exports ) Stehrer (2013) Methodology E* Net Trade VBE* Net VAiT Net Trade = Net VAiT In case 1) we use the Koopman et al. Methodology for a decomposition of each country`s gross exports. Using equation (3c) in section 3, we have calculated the value added in trade (VAiT) matrix formulized as V^BE. For instance, decomposition of gross exports of country 1 is shown in the first column, the first diagonal term measures domestic value added (DVA) in exports, which is The foreign value added (FVA) of country 2 in gross exports of country 1 is 3.33 and the FVA of country 3 in gross exports of country 1 is The sum of DVA and FVA is equal to gross exports of country 1 i.e. 20. On the first row, the second term stands for indirect exports of country 1`s value added in the exports of other countries (IV). Country 1`s value added embodied as intermediate inputs in exports of country 2 and country 3 is 3.23 and 1.48 respectively. In case 2) we use Stehrer (2013) methodology for calculating the value added in trade for each country using equation 4) which is formulized as VBE *. Note that in this case, the sum of each column of the E * matrix is the net trade balance of each country. Country 1`s trade balance with the world is -5. This is equal to country 1`s net value added trade balance as seen by the first column sum of VBE * matrix. 35

43 Chapter 3 India s Intermediate Goods Trade in the Inter-regional Value Chain: An Examination based on Trade Data and BRICS Input-Output Table Abstract: Intermediate goods play a very important role in world trade. In fact, a large part of world trade takes place in the form of intermediate goods rather than final goods. One significant feature of world trade in the past decade has been the increase in the role of China as a rapidly growing economy in the global value chains of intermediate goods trade. India s rapid growth rates have also attracted attention, but little research has been done on the nature of its integration into the value chain of trade in intermediate goods. The objective of this chapter is to fill this void in literature. Even though at the global level India s share in the trade of intermediate goods is low, if we look at the BEC (Broad Economic Categories) classifications of goods (Final, Intermediate and Primary goods), India trades the most with the world in intermediate goods. Therefore, from an Indian perspective it is worthwhile to examine the nature of India s integration into the interregional value chains of intermediate goods trade and to look at its intermediate goods trade with China which has established itself as the center of Asian Intermediate goods trade in the past decade. The novel result of this chapter is that India s dependence on China has significantly increased in the past decade. Indian computer and electronic industry stands out as the industry with highest backward linkages with China. Keywords: India, Intermediate Goods, Value Chain, Trade Statistics, BRICS 2005 Input- Output Analysis 1. Introduction Trade in intermediate inputs has been steadily growing over the last decade facilitated by fragmentation of production; increase in importance of outsourcing; and foreign direct investment. Traditional theories on international trade like Ricardian Theory of Comparative Advantage and Heckesher-Ohlin Theory that dealt primarily with trade in final goods are now being modified by scholars to accommodate the changing contours of trade in the recent decades. Studies on intermediate goods and value chain are therefore attracting more and more attention. An intermediate good is defined as an input to the production process that has itself been produced and, unlike capital, is used up in production. The difference between intermediate and capital goods lies in the latter entering as a fixed asset in the production process. Like any 36

44 primary factor (such as labor, land, or natural resources), capital is used but not used up in the production process. On the contrary, an intermediate good is used, often transformed, and incorporated in the final output. As an input, an intermediate good has itself been produced and is hence defined in contrast to a primary input. As an output, an intermediate good is used to produce other goods (or services) contrary to a final good which is consumed and can be referred to as consumption good. 23 For e.g. the production of a car require tires which requires steel and rubber; steel requires iron which requires iron ore and so and on, similarly rubber requires other inputs. In this example, car is the final product and rubber; steel; iron ore are all intermediate inputs used in its production. Intermediate inputs consist of both material goods and services. The latter can also be used as input to any sector of the economy; that is for the production of the same, or other services, as well as manufacturing goods. According to an OECD report, intermediate inputs represent 56% of total goods trade and 73% of total services trade. Trade flows are dominated not by goods that are fully consumed but by goods that are further used in the production of other goods and services (OECD, 2009). The trade of intermediate inputs has grown at an average annual rate of 6.2% for goods and 7 % for services (in volume terms) between 1995 and (Miroudot et al. 2009). Hence, intermediate goods have become an important component of world trade.. Value chain can be defined as the different stages in the production of a good which add value to the final product. Each of this stage can be called a part of the value chain. The term was popularized by Michael Porter (Porter, 1985). In recent decades there has been an increase in shifting some stages of the value chain to locations abroad. This phenomenon has been addressed in various ways. slicing up the value-added chain (Krugman 1991); offshoring ; outsourcing ; fragmentation (Jones and Kierkowski 1990 and 2001, Deardorff 2001); vertical specialization (Balassa 1967 and Hummels et al. 1998, 2001, Gonzalez 2012) and unbundling (Baldwin 2006, 2012). To refer to the geographical shifting out of stages in the value chain, two words are very commonly used in the literature on this field: Global Value Chain and Regional Value Chain. Baldwin and Gonzalez (2013) argue that the supply chains of production networks are more regional than global. Figure 1 shows the global matrix of imported inputs used in production. The rows show the sale of intermediate inputs to a column country and the columns show the purchases of intermediate inputs from a row country. It can be seen clearly from the figure that the sales of input across countries is a regional activity. For example, 23 This is a definition provided by Deardorff (2006) in his Glossary of International Economics. 37

45 Germany sells the most within Europe and China sells the most within Asia. It can be seen that very few bilateral flows are significant on a global scale. Therefore, Baldwin argues against the idea of global supply chain and supports the idea of regional production networks as shown by the three rectangles in figure 1. The global production network is marked by regional blocks, what can be called Factory Asia, Factory North America, and Factory Europe. The purpose of this chapter is to shed light on the position of India in the production networks of intermediate goods based on three observations: First, in the BEC (Broad Economic Categories) classifications of goods (Final, Intermediate and Primary goods), India trades the most with the rest of the world in intermediate goods. Second, China has significantly increased its position in the world trade of intermediate goods between 2001 and 2011 and has become the prime export base for assembly of manufactured goods that contain imported components. Third, the global production networks are more regional than global (Baldwin 2013). Based on these three observations, what can be said about India s role in the production network of intermediate goods? What kind of intermediate good trade relation does it have with China? Is Baldwin s Factory Asia argument valid in the case of India? 24 To examine these questions this chapter will use trade data and input-output analysis as the empirical tool of examination. To get a broader picture we begin by looking at India s trade with few other countries (EU, US, ASEAN, and Japan) and then eventually narrow down to China. For the purpose of this chapter, we limit ourselves to the study of goods and leave the study of India s integration into intermediate trade of services for future research even though services industry has become the most important industry in India in the past decade. 25 An analysis of trade in services sector requires data at a disaggregated level, but limitations of finding such data is the reason why we exclude services from this research. For instance, the BRICS International Input-Output Tables only includes 4 classifications of service industries: Construction, Trade and transport, other services and Public administration. Such a broad aggregation of services industry constraints the meaningful analysis of India s service industry which mainly trades with the world in software services. 24 We examine the factory Asia argument of Baldwin for the case of India in Section 4.2 of Chapter Services trade is very important for India and India s rapid growth rates have been attributed to its rapidly growing services sector. Much of this growth in exports is driven by growth in service exports. Within services, it is mainly modern services (referred to as miscellaneous services in data published by the Reserve Bank of India) that have been driving this export performance. Further decomposing miscellaneous services into software, communications, business and financial services reveals that exports are dominated by software services. Export destination of service exports is highly concentrated with U. S as the top destination (65% of total software and service exports). For more details refer Eichengreen, B. and P. Gupta (2010). 38

46 The structure of the chapter is as follows. In Section 2, we explain the motivation for research. In Section 3, we review the literature on intermediate value chain. In Section 4, we explain the data and methodology. In Section 5 and Section 6 we examine India s intermediate good trade using two different methodologies, i.e. RIEIT Trade (REITI TID) database and BRICS Input- Output table (IO) respectively. Finally, Section 7 is devoted to drawing conclusions from the analysis and identifying scope for future research 2. Motivation for Research In Asia, the attractiveness of China and ASEAN countries as a favorable manufacturing base has increased their significance in intermediate goods trade. Together these two regions account for much of the intermediate goods trade in this region. Table 1 shows the geographical distribution of world trade of goods according to production stage. In 2001, the share of China s trade as a percentage of total world trade in primary, intermediate and final goods was 1.7%, 4.4% and 10.8 % respectively. Its share of world trade was the least in primary goods and it reduced by half till On the other hand, if we look at China s trade of intermediate goods as a share of total world trade in intermediate goods, we find that it almost doubled to 8.1% in China s trade in final goods also showed a jump during the ten year period increasing to 19.5% of total world trade in final goods in the year If we look at the import of intermediate goods, there too the share of China almost doubled. The share of China in world intermediate goods imports more than doubled from 4.4% to 9% between 2001 and An interesting point to note is that the share of China in world final good import was only 5% in 2011 which is a meager amount when compared to China s share in world final good export that stood at 19.5% in the same year. From above figures we can understand that in the intermediate good category, over the 10 year period China doubled its share of both imports and exports in world trade. In the final good category it significantly increased its share only in the export of final goods. This indicates that China increased its position in the global production network mainly as the supplier of intermediate goods. This point is in tandem with the argument of Baldwin (2013) as presented in Figure 2. Figure 2 presents the global sourcing strategies matrix for domestic production at the global level. Baldwin used a comparison of 1995 and 2009 to see an increase in the position of China as a supplier of intermediate inputs shown. The rows indicate the sales by row country to the column countries. The row next to China shows an increase in sourcing to most of the countries included in the matrix. 39

47 Table 1 also shows the share of ASEAN. ASEAN including Indonesia, Malaysia, Philippines, Singapore, Thailand, Brunei (Darussalam), Cambodia, Vietnam increased its share of intermediate goods from 7.2 % in 2001 to 9.5% in The share of ASEAN export in world trade of primary goods increased a bit from 4.3% to 5.2% between 2001 and Its share in world trade of final goods reduced a little from 6.2% to 5.7% between 2001 and When we look at imports, we find that ASEAN s share of intermediate goods is also high like its exports of intermediate goods. This share increased from 7.5% to 8.2% between 2001 and The share of ASEAN imports in world trade of primary goods fell from 5.2% to 4.8%. The share of ASEAN imports in world trade of final goods increased only slightly from 3.5% to 3.9%. Thus, from the above figures we can see that ASEAN like China trades the most with the world through its export and import of intermediate goods. If we compare with China or ASEAN, India s performance in intermediate goods has been lackluster. Although India s share in global export doubled from 0.8% to 1.9% and import of intermediate goods more than doubled from 0.8% to 2.8% between 2001 and 2011, in the discussion on intermediate goods, India is often shadowed by the growth of China and ASEAN countries. Given the importance of China and ASEAN in intermediate goods trade, a lot of research has been done to examine the nature of their intermediate goods trade. However from an Indian perspective, it is important to research on India and its integration into the regional production networks for two main reasons: First, the similar economic profile of India and China in terms of population and GDP in the past few years is notable. India s rapid growth rates like China have attracted attention. But little research has been done on the nature of its integration into the interregional value chain of trade in intermediate goods. One possible explanation for this is that Indian Growth Model is characterized by IT services led industrialization. This growth model is very different from the Chinese (or East Asian Style) industrialization characterized by manufacturing export led industrialization. Figure 3 shows the structure and employment GDP in India and China, the outer circle shows the distribution of employment and GDP of India and China for the year 2000.Outer circle shows share of GDP and inner circle shows the share of employment. In India services sector contributes the most to GDP (50.4%) but employment in this sector is only 24.1% of the total work force. The contribution of manufacturing sector to GDP is only 26.4% and it employs 16.1% of the total workforce. On the other hand in China, it is the manufacturing sector that contributes the most to GDP (46%) and employs 22.5% of the total workforce. The service sector contributes to 38.9% of 40

48 GDP and employs 27.5 % of the workforce. The GDP to employment representation is more balanced in the case of China than India. In India concerns are now mounting over the lowemployment generation of India s IT led skill intensive growth model. It is argued that India is facing the situation of good export performance and jobless growth. Heavy reliance on relatively skill-intensive service sectors has created limited prospects for employment growth in India during the past two decades. An average of 13 million people can be expected to enter into India s labor force annually in the next four decades, most of whom will be unskilled workers. To absorb such a massive expansion of the workforce, India will need to develop more manufacturing sectors that are labor-intensive (UNESCAP, 2011). Therefore, a study of India s integration into trade networks is the need of the hour. Second, intermediate goods are an important part of India s trade with the world even though at the regional level India s contribution to trade is significantly low compared to China and ASEAN countries. Table 2 shows the importance of primary, intermediate and final goods in India s trade with the world. In the three BEC classifications of goods, India trades the most with the world in intermediate goods. India s intermediate exports accounted for 52% of India s trade with the world in 2001 and this increased to 60% in 2011.On the other hand, the share of India s export to world in primary goods was a meager 7% and 9% of world trade in 2001 and The share of final good exports in India s trade with the world fell from 41% to 31%.We find a similar picture of importance of intermediate goods in India s trade with the world when we look at India s imports. India imported intermediate goods the most, in both 2001 and 2011 and the share increased from 44% to 49%. Therefore, India s position in global value chain maybe weak but if we look at India s trade in goods with the world, we find it is intermediate goods which account for most of India s trade. Based on the above arguments in the following sections, we try to examine the nature of India s intermediate goods trade and its integration into the interregional value chain especially its interaction with China. 41

49 3. Literature Review on Intermediate Goods Trade and Value Chain The literature in the field of international production networks can be divided into two broad sets based on area of coverage: Global and Asian region. The first set deals with this topic as a whole with interconnection across regions: Asia, Europe, and America. The most notable and latest works in this area include the works of Baldwin and Gonzalez. Baldwin (2012) looks at concept of supply chains. He provides a general idea of global supply chains: past, present and future. According to him supply chains are as old as industry itself. He divided the evolution of supply chain into two parts - the first and second unbundling of globalization. While the first unbundling was initiated by transportation made possible by steam, the second was initiated by transmission made possible by ICT. He identifies that the world is now in the 2 nd phase of unbundling, in which some production stages previously performed in close proximity are dispersed geographically. In the previous stage of unbundling, firms had to be in close proximity to perform best but ICT revolution has helped the firms to grow over the coordination glue because now it s not too complex to manage operations across borders. 26 In his paper, he provides many schematic frameworks in which this 2 nd unbundling process is being carried out. One interesting finding of his study, relevant for our chapter, is the rise of Asia in intermediate goods trade of electronics and other related equipment. Gonzalez (2012) in the first essay of his thesis conducts a bilateral input-output analysis looking at both the country of import and destination of export. He does this by looking at backward and forward linkages of the input-output table. One of the interesting finding of his analysis is that global supply chain is more regional than global. Given the importance of China and ASEAN in intermediate goods trade, there is a wide range of literature that looks at Asian aspect of intermediate goods trade, especially China s intermediate goods trade. We club this under the second set which looks at specific region.i.e. the Asian region. Study of China and ASEAN or China and Japan are common themes under this set. The methodologies used range from trade data analysis to input-output analysis (Ha Thi Hong Van 2011), ( Makishima 2011), (Yasuhi2011), (Yamada 2004, 2006), (Wang 2004), (Wang and Uemura 2006), (Hasebe and Shrestha 2006) and (Hasegawa 2012). India even though in Asia is rarely mentioned or given a passing mention to. 26 According to Baldwin, the coordination glue began to melt from the mid-1980s because of two main reasons : The ICT revolution made it possible to coordinate complexity at distance. The vast wage differences between developed and developing nations made separation profitable. This was globalisation s 2nd unbundling some production stages previously performed in close proximity were dispersed geographically. (For more details see Baldwin 2012, Figure 3, and Page 4). 42

50 There is a scarcity of literature that examines India and its role in regional value chain. UNENSCAP (2011) and Yamashita (2012) look at India and its role in production networks using UNComtrade statistics. Athukorala (2008, 2013) evaluates India s participation in the production networks using gravity models. Most of these studies rely on trade statistics as an empirical tool of analysis and depending only on trade statistics constraints in depth understanding of trade in intermediates. Trade data is collected according to the industry of origin. Assuming the industry producing the good in one country sells it to the same industry in the other country/region. This means it does not provide information about the interindustry usage of intermediate goods across different countries/regions. However, intermediate goods and services are not only used within the same industry at higher stages of the production chain, but also by other industries. Using International Input-Output (IIO) table overcomes this weakness of the trade statistics.while trade statistics only show direct interconnection among sectors, IO analysis shows direct and indirect interconnection of inputs from various sectors.thus, an important advantage of using input-output tables is that it shows the interdependence across different industries of different countries. It was difficult to find a study that uses input-output analysis to look at the linkages of India in production networks. This is where this research aims to fill the gap in literature. This chapter will use the methodology of trade statistics as well as input-output analysis to examine the research topic. 4. Methodology and Data The objective of this chapter is to analyze the nature of India s intermediate goods trade in the interregional value chain. We are especially interested in India s intermediate goods trade with China. We use two methods: Trade Data Analysis and Input-Output Analysis. 4.1 Trade Data Analysis REITI Trade Industry database or REITI TID is a very useful way of analyzing trade in intermediate goods among many counties.the export value and import value of the countries and regions are organized by partner country (including global total), industry (13), production process (five stages/ three stages), and year. In Section 5, we look at India s intermediate goods trade with the following 3 countries and two areas China, Japan, US, ASEAN (Including Indonesia, Malaysia, Philippines, Singapore, Thailand, Brunei (Darussalam), Cambodia, Vietnam), EU27 (United Kingdom, 43

51 France, Germany, Italy, Austria, Belgium-Luxembourg, Denmark, Finland, Greece, Ireland, Netherlands, Portugal, Spain, Sweden, Bulgaria, Cyprus, Czechoslovakia, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Romania, Slovenia ). REITI TID divides each production stage into 3 categories: Primary, Intermediate and Final goods. Intermediate goods can be further divided into: Processed goods; Parts and components. 27 The above classification of goods is provided for 13 industries-foods; Textile; Pulp, Chapter and Wood; Chemicals; Oil and Coal; Stone, clay, glass and concrete products; Iron and steel, Nonferrous metals; General machinery, Electrical machinery; Household Electrical appliances; Transport equipment; Precision machinery, Toys and miscellaneous goods. The advantage of trade data as compared to I-O tables is that it helps in looking at bilateral trade patterns of intermediate goods at a disaggregated level. Furthermore, it helps in looking at changes overtime as the data for several years are available. 4.2 Input-Output Analysis There are two types of Input-Output (IO) Tables in the study of IO analysis: National IO Tables and International IO Tables (IIO). The former focuses on interactions among different sectors within an economy. The latter focuses on interactions among different sectors across different countries, the countries that are not included explicitly in the International IO Tables are clubbed into the rest of the world. We are interested in the latter.i.e. India in international IO analysis. For this, in this chapter we rely on a recently published BRICS 2005 IO table. In the next chapter, we extend this analysis using WIOD (World Input-Output Database). Institute of Developing Economies (IDE) provides the BRICs 2005 Input- Output table. This IIO table covers the following countries- Brazil, Russia, India, China, US, Japan, EU. We are interested mainly in intermediate demand and supply matrix as denoted by matrix A (Input Coefficient Matrix). A shows the input coefficient of each of the 25 industries for BRIC+3 countries. In other words, the input required to produce one unit of output. From the standard equations of IO analysis, 27 Processed goods and Parts and components can further be divided into the following items-processed goods: Food and beverages, processed, mainly for industry, Industrial supplies, n.e.s., processed, Fuels and lubricants, processed, Parts & Components: Parts and accessories of capital goods, except transport equipment; Parts and accessories of transport equipment.( REITI TID database) 44

52 X= AX + F X = (I-A) -1 F Where I stands for Identity Matrix A stands for Input coefficient Matrix F stands for Final demand. We are interested in the (I-A) -1 the Leontief Inverse Matrix which gives the linkages. Let L= (I-A) -1. Figure 4 gives the schematic representation of (I- A) matrix of 2005 BRICS IO table. I is the Identity Matrix of the same size as Matrix A with all diagonal elements =1 and all the other values = 0. The BRICS IO table is divided into four categories in matrix form based on interactions among- i, j, R, S. where i, j stand for industries and R,S stand for countries. Aij RS shows the input coefficient to be purchased by the j-th industrial sector in the country (or economy) S in the intermediate input produced in the i-th industrial sector in the country (or economy) R. There are 25 sectors in each country in the BRICS IO Table Countries= Brazil, Russia, India, China, Japan, US, EU (B, R, I, C, J, U, E) = 7 countries Aij RS = [Aij RS ] where R, S= B, R, I,C, J, U,E and i, j=1,2,3 25 The matrix in the case of R=S is domestic intermediate demand The matrix in the case of R S is to show the trade coefficient matrix from the economy R to the economy S. 25 sectors in 7 countries therefore n=25 7=175 (Size of Matrix) Input Coefficient Matrix [Aij RS ] = (domestic intermediate input coefficient + trade coefficient) square matrix From the above, we are interested in calculating the following, 45

53

54 One problem of BRICs IO table is that it is provided only for the year 2005, thus it is not possible to look at the changing interdependence overtime between India and these countries. Furthermore, it does not include ASEAN; hence it is difficult to see the interaction between India and ASEAN countries. Nevertheless, it provides a useful way to look at the intermediate goods interdependence. The drawback of both these databases is that it records the data in currency terms (US dollar) therefore making it susceptible to exchange rate fluctuations over the period under consideration. Nevertheless, the use of both these methods helps in getting a more holistic picture of intermediate goods trade. 5. India s Intermediate Goods Trade Using Trade Statistics In this section using the RIETI TID 2012 database we try to examine India s intermediate goods trade based on industry classification and geographical distribution. Table 3 shows the share of each of the 13 industry classifications of REITI in India s total intermediate good trade with the world (13 industry total).the percentage can be calculated by looking at the intermediate exports from each industry by India to the world divided by India s total export of intermediate goods (Industry total) to the world in that year. We can see that in 2001 India exported the most intermediate goods in stone, clay, glass and concrete products which stood at 26% of the total trade in intermediate goods, followed by chemicals at 19.06% and oil and coal at 16.82%. In 2011, the same industries were the main exporters but the order changed. Between 2001 and 2011 the share of most of the other industries in India s total intermediate exports fell. The only industries that showed a little bit of increase were: iron and steel and non ferrous metals, transport equipment and precision machinery. When we look at imports of intermediate goods from the world, we find the industry that imported the most in both 2001 and 2011 was iron and steel and non ferrous metals followed by chemicals. The import of all other industries fell except transport equipment: stone, clay glass and concrete products. Table 4 looks at the geographical distribution of India s trade in intermediate goods. The percentage can be calculated by dividing India s export of intermediate goods to each country with the total export of intermediate goods from India to the world. It can be seen that India exports much of its intermediate goods to EU27 followed by US. The order remained the same in both 2001 and As for imports it can be seen that India imported the most of its intermediate goods from EU 27 in 2001 but this changed in China s 47

55 share in India s intermediate goods more than doubled between 2001 and 2011 increasing from 6.29% to 13.39% of India s total import of intermediate goods import. It is also interesting to see that in the year 2011 except for China the share of all the other countries decreased. Table 5 looks at India s intermediate goods imports from China according to 13 industry classifications. The percentage can be calculated by dividing the intermediate imports from China of each industry divided by the industry total of intermediate imports from China. The share of chemicals in intermediate imports was the highest in both years amounting to 30% of India s total intermediate imports from China. In 2001, the industry that imported second most was textiles amounting to 16.64%. Between 2001 and 2011 the share of iron steel and non ferrous metals increased from 13.70% to 21.21% percent to become the industry that imported second most from China after chemicals. It is interesting to see that the share of stone, clay, glass and concrete products, which accounts for a large share of India s total intermediate imports from the world based on Table 3, has a very small share in imports from China. This goes to show that much of it comes from other countries. In 2011, at the third and fourth level were electrical machinery and general machinery at % and % of India s total imports from China, respectively. The position of these industries jumped up compared to the year It is also interesting to note that the share of transport equipment increased more than 12 times, from a mere 0.28 % to 2.46 %, making it the industry which showed the most growth intensity. To summarize, in this section using trade data we first looked at, the share of each industry in total intermediate goods trade of India (according to industry classification). We found that stone, clay, glass and concrete products; chemicals; oil and coal; iron and steel non ferrous metals were the front runners in India s intermediate goods (both exports and imports) to the world in India s exports most to EU and US but imports most from EU and China. China s position in India s imports overtook EU in 2011.The share of electrical/ machinery items i.e. general machinery, electrical machinery and transport equipment is steadily increasing in India s intermediate good import from China. This could be one reason for the jump in the position of China. 6. Input-Output Analysis of India s Intermediate Goods Trade Linkage analysis provides a very useful way to look at the interdependence of supply and demand across sectors as well as countries. The standard method for such an analysis of 48

56 interdependence is backward and forward linkages. International backward linkage can be defined as the extent to which, the industry of one country depends on the component suppliers of another country. International forward linkages determine the extent to which the industries of other countries depend of the output of the industry of a given country. Domestic backward linkages can be defined as the extent to which industry in one sector depends on the industry in another sector. Domestic forward linkages can be defined as the extent to which other industry s output depends on the output of a given industry. Based on the analysis of the previous sections, generally we can expect international forward linkages to be low for India because of India s weak position as an exporter in the region. 28 As for international backward linkages, we can expect them to be most with EU and China. Using BRICS 2005 IO table in our analysis, we are interested in the following: a. We first look at interdependence between different sectors of India and all BRIC+3 countries without differentiating the country of origin. In this part, the intensity of backward linkage (BL) is measured with an index known as the Power of dispersion. Similarly, the intensity of forward linkage (FL) is measured with an index known as the Sensitivity of Dispersion. An index value of more than one means that the industry has a power of dispersion/sensitivity of dispersion greater than the average of all industries. However, this total figure for power of dispersion includes both domestic and external backward linkages. To look at just the external BL/FL we subtract domestic BL/FL from the total to calculate the External Backward Linkage (EBL) and External Forward Linkage (EFL). b. Next, we look at interdependence between different sectors of India and each of the BRIC+3 countries. For this, we look at India s industry wise backward and forward linkage to each country. This can be calculated as the sum of column (for backward linkage) and rows (for forward linkage) of the Leontief inverse matrix. For the first part, we tried to identify the industries within the top 4 rank in power of dispersion and sensitivity of dispersion index. Our analysis shows the following: When we look at the total figure of power of dispersion of India as can be seen in Table 6, it is the highest for computer and electronic equipment, followed by, other electrical equipment, textile, leather, and the products thereof and finally food, beverage and tobacco. 28 The result shown on Table 7 confirms this point. 49

57 In other words, these industries depend more on component suppliers when compared to other industries. As can be seen from Table 6, external backward linkages were also the most for computer and electronic equipment, other electrical equipment, followed by transport equipment and chemical products in that order. That is, relative to the other industries, these industries purchase most inputs from the industries of BRIC+ 3 countries. Textile, leather, and the products thereof; & food, beverage and tobacco do not figure high in external backward linkage implying that the backward linkages in these industries are more domestic than external. The ranking of domestic BL is also shown for reference. The table shows that out of all industries computer and electronic equipment clearly tops the average of all other industries in terms of both domestic and external backward linkages. Table 7 shows the sensitivity of dispersion of top 4 industries. When we look at sensitivity of dispersion, the index is highest for trade and transport followed by other services, chemical products, metal and metal products. In other words, the influence of these sectors on the economy as suppliers is higher than the average of all industries. Just like in the case of external BL, we calculate external forward linkage (FL). The ranking of the top four industries is in the following order: chemical products, trade and transport, other mining, petroleum and petro products. The contribution of these four industries to the index of sensitivity of dispersion was the most. That is, relative to the average of the other industries, these industries supply more of their output to industries of BRIC+ 3 countries. Under the category of manufacturing intermediate goods, it is the chemical products industry tops the list. In the second part of this section as mentioned above, we look at country specific linkages between different sectors of India and each country of BRICS as shown in Table 8. At the overall level, we see that India has maximum backward linkages with EU followed by China. With both these partners, among all the industries, the backward linkages are the highest in computers and electronic equipment followed by, other electrical equipment, industrial machinery, transport equipment and chemical products. An analysis of the external forward linkages of each Indian industry to each BRIC+ 3countries, reveals a very weak position of India in supply of intermediate goods. The only industry in which India has significant forward linkages is other mining industry. These 50

58 linkages are highest with China and stand at 3.1%, for all the other countries the figure is less than 0.4%. 29 Thus, the important findings based on tables, Table 6, 7 and 8 can be summarized as follows: EBL are highest in the following industries- computer and electronic equipment, other electrical equipment, transport equipment and chemical products. These external linkages are most with EU followed by China for all these industries. 7. Conclusion Based on the results of our trade data analysis and international input-output analysis of value chains of the Indian economy, we can draw the following conclusions: 1. Chemicals stand out as one of the top industry for India based on both trade statistics (Export / Import/ Import from China) and IO analysis (Rank 4 in our analysis of power of dispersion). In other words, not only is the share of this industry high in India s trade basket but also the linkages of this industry are high in both domestic and external linkages. 2. The share of iron and steel in India s import basket was the highest, but it does not figure among top industries with linkages. This implies this industry has weak backward linkages with other industries. This could be because the inputs are produced and consumed in the industry of origin and industry of import, respectively. 3. Out of all the industries, the power of dispersion (Index of backward linkage) was the highest for computer and electronic equipment, other electronic equipment, textiles and food in that order. Even if we exclude the domestic inter industry linkages to look at just the external power of dispersion, computer and electronic equipment, other electronic equipment; continue to hold their position. 4. A country wise look at India s external backward linkages shows that EU takes the top place and China takes the second place. Among the industries with the linkages with these countries, on the top figure, computer and electronic equipment, other electronic equipment followed by industrial machinery, transport equipment, and chemicals. 5. Overall backward linkages of the computer and electrical equipment and other machinery industries are the highest, in spite of the fact that the shares of those industries are not so high in India s trade basket. The result of the analysis is 29 Given the weak forward linkages, these results were are not presented but are available on request. 51

59 consistent with Baldwin s argument, 21st century trade is concentrated in relatively few sectors. Electrical machinery and electronics take the lion s share of the level and the growth in the 1990s (Baldwin 2012, Pg.6). As a first step in research, this chapter tried to identify India s position in inter regional value chain of intermediate goods using the methodology of Input-Output analysis as well as trade data. We found India is lagging far behind the other Asian economies in benefitting from the new form of fragmentation based production network trade, this can be seen from our results on forward linkages. Based on trade data analysis our results confirm the previous studies in this area that India s participation in regional production networks is low (UNESCAP 2011, Athukorala 2013). India continues to trade with the world mostly in resource based intermediate goods like stone, clay, glass and concrete products; chemicals; oil and coal; iron and steel non ferrous metals (both export and import). This result is in synch with the UNESCAP (2011) study which compared India and China value chain activity to find India s merchandize trade to be resource based while China is moving up the value chain ladder from low skill intensive electronic products to high skill intensive electronic products. In this context, the recent trend of rising wages in China opens up an opportunity for India to cash in on its low cost labor advantage as well as population boom and formulate trade policies for better integration into these production networks. One novel conclusion of this chapter is the importance of China in India s international value chain. The past researches in this field have only used trade statistics which fail to give a complete picture as they do not show interdependencies among different industries of different countries. As a result of using the IO analysis, in our study we were able to observe the high level of linkages between India and China in computer and electronics industry which otherwise were not visible if we look only at the trade data analysis result. This result of the chapter confirms Baldwin and Gonzalez (2013) argument that the supply chain is more regional than global. Based on trade data analysis comparison of 2001 and 2011, we found that trade with China especially imports from China has increased significantly. Using trade data, chemicals or iron and steel, are the top Indian imports from China. However, if we look at backward linkages using input-output analysis, linkages with China are the most in computer and electronic equipment related industries. This conclusion has policy implications for India s trade with China. 52

60 Figures and Tables Figure 1 Global Matrix of Imported Inputs to Produce Baldwin and Gonzalez (2013), Figure 13, Page 19. Bilateral purchases of intermediates by row nation from column nation as % of all flows in WIOD data base; flows under 0.3% set to zero. 53