Compilation of a regionally extended inter-country input-output table and its application to global value chain analyses

Meng, Bo; Yamano, Norihoko Article Compilation of a regionally extended inter-country inputoutput table and its application to global value chain analyses Journal of Economic Structures Provided in Cooperation with: Pan-Pacific Association of Input-Output Studies (PAPAIOS) Suggested Citation: Meng, Bo; Yamano, Norihoko (2017) : Compilation of a regionally extended intercountry input-output table and its application to global value chain analyses, Journal of Economic Structures, ISSN 2193-2409, Springer, Heidelberg, Vol. 6, Iss. 23, pp. 1-38, https://doi.org/10.1186/s40008-017-0081-z This Version is available at: https://hdl.handle.net/10419/194889 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Compilation ofa regionally extended inter‑country input–output table andits application toglobal value chain analyses Bo Meng1* and Norihoko Yamano2 1 Background As one of the most useful tools for studying international production networks, intercountry input–output (ICIO) tables provide detailed information and thus have been widely used in both economic and environmental analyses (see Miller and Blair 2009; Murray and Wood 2010), especially in areas related to today’s global value chains (GVCs) (see WTO-IDE 2011; OECD-WTO 2013; OECD-WTO-World Bank Group 2014; Koopman etal. 2014 (KWW); and Meng etal. 2015). If the focus of analysis is at only the country or inter-country level, the existing ICIO tables may be sufficient. In practice, however, many economic policy needs are at a country’s domestic regional level, while linking with the world market. For example, questions like how the 2008 financial crisis damaged the economy of Chinese Guandong Abstract Studies on the rise of global value chains (GVCs) have attracted a great deal of interest in the recent economics literature. However, due to statistical and methodological challenges, most existing researches ignore domestic regional heterogeneity in assessing the impact of joining GVCs. GVCs are supported not only directly by domestic regions that export goods and services to the world market, but also indirectly by other domestic regions that provide parts, components, and intermediate services to final exporting regions. To better understand the nature of a country’s position and degree of participation in GVCs, we need to fully examine the role of individual domestic regions. Understanding the domestic components of GVCs is especially important for larger economies such as China, the USA, India, and Japan, where there may be large variations in economic scale, geography of manufacturing, and development stages at the domestic regional level. This paper proposes a new framework for measuring domestic linkages to global value chains. This framework measures domestic linkages by endogenously embedding a target country’s (e.g., China or Japan) domestic interregional input–output tables into the OECD inter-country input–output model. Using this framework, we can more clearly understand how global production is fragmented and extended internationally and domestically. Keywords: Input–output, Global value chains, Regional heterogeneity Open Access © The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. RESEARCH Meng and Yamano Economic Structures (2017) 6:23 DOI 10.1186/s40008‑017‑0081‑z *Correspondence: [email protected].jp 1 Development Studies Center, Institute of Developing Economies, JETRO, 3-2-2, Wakaba, Mihama-ku, Chiba 2618545, Japan Full list of author information is available at the end of the article Page 2 of 38 Meng and Yamano Economic Structures (2017) 6:23 province through various channels of GVCs1 or how the 2011 Great East Japan Earthquake impacted Korea’s semiconductor industry through GVCs2 significantly challenge the existing ICIO approach. The conventional way of using ICIO-based economic models (including computable general equilibrium models) to answer the first question is to first evaluate the inter-country impact and then to conduct a top-down analysis at the domestic regional level. The response to the second question is to first conduct a bottom-up analysis to evaluate the impact of a specific region on the whole country and then to measure the inter-country impact. For both approaches, in the process of evaluating the inter-country impact based on the ICIO table, the target country is treated as a single entry (an economic point), without any information about its domestic regional heterogeneity. Namely, there is implicitly a strong assumption that all domestic regions have the same production function as the national average. This may potentially lead to a large estimation bias in economic analyses for the target country if it has domestic regions with various heterogeneities in terms of regional economic endowments, geographic locations, developmental stages, industrial structures, and foreign independency. Another example is that, in the USA, much of the chemical industry is concentrated in the South with large amounts of relatively clean natural gas-fired power plants. What remains of the steel industry is largely concentrated in the Upper Midwestern states, where coal-fired power plants dominate. If we use ICIO to estimate embodied emissions in US exports through domestic supply chains, an overestimation will likely occur for chemical exports, and an underestimation may occur for steel exports if more chemical exports are from the South and more steel exports are from the Upper Midwestern states. This is a crucial issue with the ICIO-type approach, which needs to be carefully treated in both economic and environmental analyses. One option to overcome this problem is to construct a regionally extended intercountry input–output (REXICIO) table containing the target country’s domestic regional information. This can directly provide information about linkages between a target country’s domestic production networks and international production networks. Some efforts have been reported in the literature. For example, Meng etal. (2013b) embedded China’s domestic inter-regional IO (IRIO) table into the world input–output table (WIOT) using a linear programing method. Their empirical results show that using the newly extended WIOT can both reduce the estimation discrepancy of China’s bilateral trade in value added and also elucidate the position and the degree of participation of China’s domestic regions in global value chains which cannot be explicitly measured by using China’s domestic IO table or the WIOT in isolation. This new WIOT has also been applied to environmental analysis (see Pei etal. 2016). Similar works can also be found in Cherubini and Los (2013), who link Italy’s IRIO table to the WIOT, and then illustrate Italy’s regional employment patterns in a globalizing world. Also, Dietzenbacher etal. (2013) linked Brazil’s IRIO IO 1 “An estimated 600,000 migrant workers have left China’s southern Guangdong Province due to unemployment in 2008 after the worldwide financial crisis hit the region” (China Daily, www.chinadaily.com.cn/business/2009-01/08/content_7379756.htm). 2 “Japan earthquake to affect several major IT industries of Korea” (DIGITIMES, by Betty Shyu, (http://www.digitimes. com/news/a20110322VL203.html). Page 3 of 38 Meng and Yamano Economic Structures (2017) 6:23 table into the WIOT and evaluated the role of Brazilian regions in the global value chain. Another ambitious attempt is to link more countries’ domestic IRIO IO tables to existing ICIO tables [see Meng etal. (2013a) and Inomata and Meng (2013), who links Japan, China, and Korea’s domestic IRIO tables into the Asian International IO tables]. The aim of this paper is to develop a more consistent and flexible method for embedding a target country’s domestic IRIO table into an existing ICIO table. The existing efforts mentioned above can be considered a good starting point. However, some important problems in the existing works remain unsolved. One is consistency of terms used for valuation in the IO tables. For example, the official WIOT uses basic prices, while China’s IRIO table uses producer prices. Without appropriately adjusting China’s IRIO from producer’s prices to basic prices in advance, potential discrepancies may arise to some extent in the process of linking these two tables. This is mainly because taxation and transportation margin at the product level may be different across domestic regions or sectors. In addition, the import data in China’s IRIO table are at CIF prices, which also should be adjusted before linking with an ICIO table since tariff, import duty, and international transportation margins and insurance may vary across countries of origin at the product level. The other challenge is how to improve the reliability of the target country’s regional trade data (exports and imports for goods and services). These data provide very important information to be used in allocating the international trade flow between the target country’s domestic regions and other countries. However, domestic regional customs data may not provide true domestic destination information of imported goods and domestic origin information of exported goods. For example, officially published Japanese regional customs data are based on the reporter’s location information rather than that of producer and user locations. Therefore, when using a target country’s domestic regional customs data, careful treatment and adjustment should be done in advance. Another problem is the control and evaluation of estimation errors. Meng etal. (2013a) and Inomata and Meng (2013) use the so-called split method to link target countries’ domestic IRIO table with an ICIO table. This method keeps the balance of the existing ICIO column-wise, but without detailed treatments of row-wise estimation discrepancies. Meng etal. (2013b) use a linear programing method to do similar linking work resulting in a balanced ICIO table with a target country’s IRIO table embedded. However, the selection of constraint equations in the linear programing method is not unique. In order to get a converged result with the smallest discrepancy level, personal experiences play a very important role in their estimation. In order to overcome the above limitations, this paper introduces some new variables for adjusting the valuation format between different tables to be linked. We also improve the reliability of a target country’s domestic regional trade data by using the OECD enduse category for customs information. In the Sect.2, we establish a stylized framework to compile the REXICIO table, and show how to use this table to study GVCs. In Sect.3, we use China and Japan as an example to show the application results of the REXICIO tables in GVC research. Section4 gives conclusion. Page 4 of 38 Meng and Yamano Economic Structures (2017) 6:23 2 Methods 2.1 How toembed a specific country’s domestic inter‑regional IO table intoan existing inter‑country IO table 2.1.1 Framework ofthe regionally extended inter‑country IO table anddata configuration For ease of explanation, we consider a two-country case where the target country has two domestic regions and two sectors for each region. The existing data that can be used to construct the REXICIO table for a target country are shown below: 1. The target country’s domestic IRIO table with separate import row vector (or matrix) and export column vector (Table1). 2. A closed ICIO table including the target country as an endogenous part at basic prices (Table2). 3. Domestic regional export data by sector and by country of destination at FOB prices and domestic regional import data by sector and by country of origin at the CIF prices from the target country’s customs statistics. 4. If the table of item (1) is not at basic prices, then relevant tax, domestic and international transportation margins (including insurance), and tariff information (including import duty and commodity tax) for the target country at both the regional and product level should be available or can be estimated. For simplicity, three final demand items (household consumption, government consumption, and capital formation including change of inventory) and one value-added item are considered in the format. The format of the REXICIO table is shown in Table3. We can see that the domestic IRIO table of the target country (country 1) has been embedded in an ICIO framework. The notation used to specify IO-related variables in the paper is given in Table4. The country, region, and sector dimension configuration used in this paper are shown below. For sector, i,j∈{1, 2, ...,ns}, where i and j, respectively, represent the sectors allocated row-wise and column-wise in an IO table, and ns represents the number of sectors. For country, R,S∈{1, 2, ...,T,...,G}, where T represents the target country that needs to be embedded in the ICIO table, R and S, respectively, represent the countries of origin and destination, and G represents the number of countries. For domestic region, r ,s∈T  1, 2, ...,g , where r and s, respectively, represent the target country T’s domestic regions of origin and destination, and g represents the number of regions. For the final demand item, k ∈  1, 2, ...,nf , where nf represents the number of final demand items. For simplicity, we consider only one value-added item. In addition, the regional export and import data taken from the customs statistics for trade in goods are further separated into three main categories (intermediate goods, Page 5 of 38 Meng and Yamano Economic Structures (2017) 6:23 Table 1 Layout ofthe target country’s domestic IRIO table A target country’s domestic IRIO table Intermediate demand Final demand Export out‑ sidetarget country Total output Region 1 Region 2 Region 1 Region 2 Sector 1 Sector 2 Sector 1 Sector 2 Household consumption Government consumption Capital formation Household consump‑ tion Government consumption Capital formation Region 1 Sector 1 Sector 2 Region 2 Sector 1 Sector 2 Import outside target country Value added Total input Page 6 of 38 Meng and Yamano Economic Structures (2017) 6:23 Table 2 Layout ofan ICIO table An ICIO table Intermediate demand Final demand Total output Country 1 Country 2 Country 1 Country 2 Sector 1 Sector 2 Sector 1 Sector 2 Household consumption Government consumption Capital forma‑ tion Household consumption Government consumption Capital forma‑ tion Country 1 Sector 1 Sector 2 Country 2 Sector 1 Sector 2 Value added Total input Page 7 of 38 Meng and Yamano Economic Structures (2017) 6:23 Table 3 Layout ofan REXICIO table A Regionally extended ICIO table Intermediate demand Final demand Total output Country 1’s Region 1 Country 1’s Region 2 Country 2 Country 1’s Region 1 Country 1’s Region 2 Country 2 Sector 1 Sector 2 Sector 1 Sector 2 House‑ hold consump‑ tion Govern‑ ment consump‑ tion Capital forma‑ tion House‑ hold consump‑ tion Govern‑ ment consump‑ tion Capital forma‑ tion House‑ hold consump‑ tion Govern‑ ment consump‑ tion Capital forma‑ tion Country 1’s Region 1 Sector 1 A1 C1 A2 C2 Sector 2 Country 1’s Region 2 Sector 1 Sector 2 Country 2 Sector 1 B1 B2 Sector 2 Value added Total input Page 8 of 38 Meng and Yamano Economic Structures (2017) 6:23 household consumption goods, and capital goods) using the OECD end-use categories (mainly based on the Broad Economic Categories (BEC) defined by the United Nations Statistics Division). The regional export data for services without information about destination country can be obtained from the domestic IRIO table. If the import matrix is not available in the domestic IRIO table, regional import data for services are better obtained from other official statistics (while there is no relevant way to separate trade in services into end-use categories). The notation used to express regional export and import data is shown below. mxRs i Region s’s imports of the target country from country R for the intermediate good i. myRs ik Region s’s imports in the target country from country R for the final good i. exrS i Region r’s exports from the target country to country S for the intermediate good i. eyrS ik Region r’s exports from the target country to country S for the final good i. ms i Region s’s imports of service i. If the domestic IRIO table for the target country is at the producer’s price (e.g., China and Japan’s cases), the following supplementary information will be helpful for more reliable estimation of the REXICIO table. αrs ij(k) Adjustment item (rate) for transferring the domestic IRIO table from producer’s price to basic price. The first approximation is to assume that αrs ij(k) =α T ij(k) , where j is the intermediate use and k is the final use. βRs i Adjustment item of international transportation freight and insurance for transferring the target country’s regional imports from CIF price to FOB price (as a percentage share of CIF price). The first approximation is to assume that βRs i =β RT i. τRT i Target country’s import duty and commodity tax (as a percentage share of CIF price). γR i Target country’s trade partner’s domestic transportation and trade margin on exports (as a percentage share of FOB price). Table 4 Variable andparameter definitions exrS p and eyrS pk used in the paper are from customs data rather than IO tables. They represent target country’s domestic region r’s exports to country S for good p by end‑use categories (ex: intermediate goods; ey: final goods) ICIO table (known) Domestic IRIO table (known) REXICIO table ( R,S�= T ) (unknown) Transaction of intermediate products ¯ xRS ij xdrs ij x RS ij =¯ x RS ij ,x rS ij ,x Rs ij ,x rs ij Transaction of final demand products ¯ yRS ik ydrs ik y RS ik =¯ y RS ik ,y rS ik ,y Rs ik ,y rs ik Import row vector (or matrix) MXs j , MY s k ( mxs ij , my s ik ) Export column vector EXr i Output ¯ XR i XDr i X R i = ¯ X R i ,X r i Value added ¯ VS j VDs j VS j = ¯ V S j ,V s j Column-sum of final demand ¯ YS k YDs k YS k = ¯ Y S k, Y s k Page 15 of 38 Meng and Yamano Economic Structures (2017) 6:23 subject to The balancing conditions row-wise (row control totals) in terms of the target country’s regional exports are given by Eqs.(30) and (30a) for intermediate goods and services, respectively, and by Eqs.(32) and (32a) for final goods and services, respectively. (29) Minimize F3=  r  S i j xrS ij −ˆ xrS ij  2 ˆ xrS ij +  r  S i k yrS ik −ˆyrS ik 2 ˆyrS ik , (30)  S  j xrS pj =  S  j ¯ xTS pj ·SexrS p·  1−γr p   r  SexrS p·  1−γr p , (30a)  S  j xrS qj =  S  j ¯ xTS qj · EXr q  rEXr q , (31)  r  i x rS ij =  i ¯ x TS ij , (32)  S  k yrS pk =  S  k ¯yTS pk ·SkeyrS pk ·  1−γr p   r  S  keyrS pk ·  1−γr p , (32a)  S  k yrS qk =  S  k ¯yTS qk · EXr q  rEXr q , (33)  r  i y rS ik =  i ¯y TS ik , (34)  r x rS ij =¯ x TS ij , (34a)  j  r x rS ij =  j ¯ x TS ij ,, (35)  r y rS ik =¯y TS ik , (35a)  r  k y rS ik =  k ¯y TS ik , (36)  S  j x rS ij +  S  k y rS ik +  s  j x rs ij +  s  k y rs ik =X r i . Page 16 of 38 Meng and Yamano Economic Structures (2017) 6:23 Equations(31) and (33) represent the balancing conditions column-wise (column control total) for the same block. Equations(34) and (35) give the individual cell control inside the transaction blocks for intermediate and final products, respectively. Equations(34a) and (35a) are the relaxed balancing conditions from Eqs.(34) and (35). There is no need to give a column-balancing condition for the whole REXICIO table in terms of the target country, since according to Eqs.(12), (24), and (31), the column balance has been guaranteed (self-evidenced). However, there is no guarantee for row balance across the whole table in terms of the target country. For this reason, we use Eq.(36) to provide the row-balancing condition. Note that for Eqs.(30)–(36), S�= T. Up to this point, we have shown how a country’s domestic IRIO table can be consistently embedded into an ICIO framework by using linear programming models for different blocks, one by one. With sufficient calculation capacity for more systematic work, we can aggregate all blocks together and solve the linear programming problems at the same time. To maintain the consistency of bilateral trade balance, we must add the following constraints on the entire linear programming problem. Thus, we have the following linear programming problem for estimating the REXICIO table: subject to the equation set A (Eqs.11, 12, 15, 16, 19–21, 23–26, 30–33, 36–38) and subject to all possible combinations in the equation set B (13, 14, 17, 18, 27, 27a, 28, 28a, 34, 34a, 35, 35a) using a grid-search method. The equation set A gives the minimum necessary constrains that can guarantee a solution of the above linear programing with demand–supply equilibria for all rows and columns in the REXICIO table. In other words, when aggregating the target country’s inter-regional parts of the REXICIO table, we can get an ICIO table which is just the original ICIO table under these constrains.4 2.2 An application ofthe regionally extended ICIO table intracing value added ingross exports forboth domestic andglobal value chains 2.2.1 Measuring bilateral trade invalue added In a closed international IO framework (for simplicity, number of countries=G, number of sectors=N, number of final demand items=1, number of value-added items=1), the world’s total GDP can be given as (37)  r  j x rS ij −  r  j x Sr ij =  j ¯ x TS ij −  j ¯ x ST ij ,(S�= T) , (38)  r  k y rS ik −  r  k y Sr ik =  k ¯y TS ik −  k ¯y ST ik ,(S�= T) . (39) Minimize F=F1+F2+F3, 4 When just using the equation set A, there is no guarantee that the target function F (Eq.39) can get the best solution since there are many additional constraints (set B) remaining. Here, we can propose a kind of grid-search method to add all possible combinations in set B to solve the above linear programing problem and obtain the minimum F. Note that there are 12 equations in set B, so the possible number of combinations is C1 12 +C 2 12 +···+C 12 12 = 4095. (40) GDP =diag(V)·(I−A)−1·Y=diag(V)·B·Y Page 17 of 38 Meng and Yamano Economic Structures (2017) 6:23 where GDP =  GDP1, GDP2,..., GDPG t , V =  V 1 ,V 2 ,...,V G , A =    A 11 ··· A 1G . . ..... . . A G1 ··· A GG    and B =    B 11 ··· B 1G . . ..... . . B G1 ··· B GG    , Y =    Y 11 . . . Y G1    +···+    Y 1G . . . Y GG    . GDPRis a N*1 column vector representing country R’s GDP by sector; VR in this section is a 1*N row vector representing country R’s value-added ratio (the share of value added in total input) by sector; ARS is a N*N matrix showing intermediate input coefficients (the share of intermediate imports coming from country R in country S’s total input); BRS is a N*N sub-matrix of the international Leontief inverse; and YRS is a N*1 column vector representing country S’s final demand on products produced in country R. Following the concept proposed by Johnson and Noguera (Johnson and Noguera 2012), country R’s value-added export to country S ( TiVARS ) is defined as the value added induced in country R by country S’s final demand: when applying the above TiVA concept to our REXICIO table, the region-by-region, region-by-country, and country-by-region value-added exports can be easily measured. This can help us understand how value added is created across both regional and national borders. 2.2.2 Tracing value added ingross exports forboth domestic andglobal value chains To illustrate the performance of a country’s domestic regions in GVCs, we apply the KWW gross export decomposition method to our REXICIO system. Using this method, we can see how GVCs are fragmented and extended inside a specific country. The KWW decomposition method is shown in Fig.1. A country’s total exports in gross terms can be decomposed into three parts: value-added exports (VT), domestic content in intermediate exports that ultimately return home (VS1*), and foreign content (VS). Every part at this stage can be further decomposed into three more parts. VT yields (1) domestic value added (DV) in direct final goods exports, (2) DV in intermediate exports absorbed by direct importers, and (3) DV in intermediates re-exported to third countries. VS1* is separated into (4) DV in intermediates that return via final imports, (5) DV in intermediates that return via intermediate imports, and (6) double-counted intermediate exports produced at home. VS is further decomposed into (7) foreign value added (FV) in final goods exports, (8) FV in intermediate goods exports, and (9) doublecounted intermediate exports produced abroad. (41) ( 0, ..., TiVARS,...0)t=diag�0, ...,V R,...0�   B 11 ··· B 1G . . ..... . . BG1 ··· BGG       Y 1S . . . YGS    TiVARS =diag � VR �� BR1Y1S +BR2Y2S +...+BRGYGS � Page 18 of 38 Meng and Yamano Economic Structures (2017) 6:23 When using the notation in terms of IO techniques, the KWW decomposition method in an international IO system with n sectors and G countries can be given as follows: Here, u is a row vector of 1’s, Es represents country S’s export by sector, and Vs is the diagonal matrix as constructed by country S’s sectoral value-added rate (non-diagonal elements are given by 0). BSR is the sub-matrix of the international Leontief inverse representing the induced output by way of international production networks in country S when there is a single-unit increase in the final demand in country R. YSR represents country R’s final demand for goods and services produced in country S. ARS is the international intermediate input coefficient representing the amount, by sector, of intermediate inputs (imports) coming from country R when country S produces one unit of output. (42) uE S∗ =VT S∗ +VS1 S∗ +VS S∗ =   VS G � R�=S BSS YSR +VS G � R�=S BSS YSR +VS G � R�=S G � O�=S,R BSRYRO   =   VS G � R�=S BSRYRS +VS G � R�=S BSRARS (I−ASS )−1YSS    +VS G � R�=S BSRARS (I−ASS )−1ES∗ =   G � O�=S G � R�=S VOBOS YSR + G � O�=S G � R�=S VOBOS ASR(I−ARR)−1YRR    + G � O �= S VOBOS ASR G � R �= S (I−ARR)−1ER∗. Fig. 1 KWW’s gross export accounting system Source: Koopman et al. (KWW 2014) Page 19 of 38 Meng and Yamano Economic Structures (2017) 6:23 Since the REXICIO table includes both domestic regions and foreign countries, we must distinguish between these dimensions in our notation. For simplicity, we use R, S, and O to represent countries and r, s, and o to represent domestic regions. In the REXICIO system, the number of countries is given by G and the number of regions by g. When focusing on the decomposition of VT as shown above and using the notation shown for country and region, the extended decomposition incorporating a country’s domestic regions into an inter-country IO system can be given as follows: Here, VTs* represents region s’s value-added exports and outflows. In particular, outflows mean domestic trade flows across regions. The first term on the right side of Eq.(43) represents region s’s value-added outflow in GVCs by domestic segment. This term includes three parts. The first represents region s’s value added in direct final goods outflow; the second shows region s’s value added in intermediate outflows absorbed by direct domestic demander, and the third is region s’s value added in intermediates reshipped to third domestic regions. The second term on the right side of Eq.(43) represents region s’s value added in intermediates re-shipped to third domestic regions by way of international segments of GVCs. The third term represents region s’s value-added exports by way of international segments of GVCs. This term can be further separated into two parts. The first is region s’s value added in intermediates exports absorbed by direct international importers, and the second is region s’s value added in intermediates re-exported to third countries. The final term on the right side of Eq.(43) shows region s’s value-added exports by way of domestic segments of GVCs. The first part in this term represents region s’s value added in direct final goods exports, and the second represents region s’s value added in intermediates re-exported to third countries. Using the extended KWW decomposition technique in an REXICIO framework, the measurement of GVCs can be divided into international and domestic segments. This framework can help us understand how, and by what routes, a country’s domestic regions engage in GVCs. The method used to distinguish the domestic and international segments in the above decomposition method is based on block matrixes in the Leontief inverse used. If the notation in the block matrix involves only domestic regions, we consider the value added induced by this block matrix to be achieved by the domestic segment of GVCs. For the other block matrices in which a country notation such as R, S, or T are involved, we consider the value added induced by these block matrices to come through the international segment of GVCs. However, Bsr may not exactly represent a pure domestic segment of GVCs, since this inter-regional block matrix (Bsr) is obtained from the large matrix of the Leontief inverse based on the extended table. If there are no international segments in the REXICIO table, we cannot have Bsr. To more clearly define the pure domestic and pure (43) VT s∗=    Vs g � r�=s BssYsr +Vs g � r�=s BsrYrr +Vs g � r�=s g � o�=s,r BsrYro    +Vs G � R g � o�=s BsRYRo . =    Vs G � R BsRYRR +Vs G � R G � O�=R BsRYRO   +   Vs G � R BssYsR +Vs g � r�=s G � O BsrYrO    Page 20 of 38 Meng and Yamano Economic Structures (2017) 6:23 international segments, we introduce a block matrix of Bsr d to the above extended KWW decomposition form. This block matrix is from the large matrix of the Leontief inverse based on the domestic inter-regional IO table. The difference between Bsr and Bsr d is the international feedback effect (see Miller and Blair 1985) between the domestic and international segments in GVCs. Using this definition, we can rewrite Eq.43 in the following form: The first row of the equation shows the value-added outflow achieved by the pure domestic segment of GVCs (VOD) in region s. The second row shows region s’s valueadded outflow by way of the pure international segment of GVCs (VOI). The third row represents region s’s value-added exports through the pure international segment of GVCs (VED), while the final row shows region s’s value-added exports by way of the pure domestic segment of GVCs (VEI). A detailed description of each part is given in Fig.2. 2.3 Data used We use the following data to embed China’s 2007 domestic IRIO data into the OECD ICIO framework; at the same time, we also do the similar work using Japan’s 2005 domestic IRIO data as well: 1. China’s domestic inter-regional IO (CIRIO) table for 2007. 2. Japan’s domestic inter-regional IO (JIRIO) table for 2005. 3. The OECD inter-country input–output table (ICIO) for 2005 and 2007. 4. China and Japan’s customs import and export statistics at the provincial level for 2007 and 2005, respectively. 5. China and Japan’s domestic transportation margin and insurance from the by-product information obtained from IDE-JETRO’s 2005 Asian International Input–Output Table project. (44) VT s∗=    Vs g � r�=s Bd ssYsr +Vs g � r�=s Bd srYrr +Vs g � r�=s g � o�=s,r Bd srYro    +   Vs g � r�=s�Bss −Bd ss�Ysr +Vs g � r�=s�Bsr −Bd sr�Yrr +Vs g � r�=s g � o�=s,r�Bsr −Bd sr�Yro    +Vs G � R g � o�=s BsRYRo =   Vs G � R BsRYRR +Vs G � R G � O�=R BsRYRO   +   Vs G � R�Bss −Bd ss�YsR +Vs g � r�=s G � O�Bsr −Bd sr�YrO   +   Vs G � R Bd ssYsR +Vs g � r�=s G � O Bd srYrO   . Page 21 of 38 Meng and Yamano Economic Structures (2017) 6:23 Regional value-added outflow and export (VT) Regional value-added export (VE) Regional value-added outflow (VO) By “pure” domestic segments of GVC (VOD) By “pure” international segments of GVC (VOI) By “pure” International segments of GVC (VEI) By “pure” domestic segments of GVC (VED) Regional value-added in direct final goods outflow (VOD1) Regional value-added in intermediates outflows absorbed by direct domestic demander (VOD2) Regional value-added in intermediates re-shipped to third domestic regions (VOD3) Region value-added in intermediates re-shipped to third domestic regions by the way of international segments of GVC (VOI1) Regional value-added in intermediates exports absorbed by direct international importers (VEI1) Regional value-added in intermediates re-exported to third countries (VEI2) Regional value-added in direct final goods exports (VED1) Regional value-added in intermediates re-exported to third countries (VED2) Region value-added outflow by feedback effects between domestic and international segments of GVC (VOI2) Region value-added export by feedback effects between domestic and international segments of GVC (VEI3) Fig. 2 Decomposition of regional value-added outflow and export by GVC routes Page 22 of 38 Meng and Yamano Economic Structures (2017) 6:23 6. International transportation freight and insurance rate estimated by using the CEPII’s 2005 and 2007 BACI database (including international trade data at both CIF and FOB prices). The 2007 CIRIO tables are compiled by China’s State Information Center (see Zhang and Qi 2012). The most detailed table has 8 domestic regions and 29 sectors at the producer’s price with a stand-alone row vector for imports. The main data source when compiling CIRIO tables is from every province’s regional IO table. One problem is that Tibet has no original IO table for 2007. Fortunately, the value-added data and final demand data by industry for Tibet can be obtained from officially published statistics. This information has been added to the southwest region in order to keep consistency with the officially published national value added and final demand information when doing the final balancing work (total inputs should equal total outputs by sector) for the whole inter-regional IO table. In addition, this treatment for Tibet causes very limited bias on the analytical results, since Tibet’s GRP as a share in China’s total GDP is just 0.137% for 2007. The 2005 JIRIO table is from the Ministry of Economy, Trade and Industry, Japan (METI).5 This table covers 9 domestic regions with 53 sector classifications at the producer’s price. Note that the import information is a stand-alone column vector in the JIRIO table, not a separate matrix (the information of imports by product is available, but no information about how many imported products are used in which industry). To get the transaction flow of pure domestic products across sectors and regions, we use the following method to remove the imported parts from current transaction flows in the JIRIO table. In practice, when converting a competitive-type national IO table (an IO table without a separated import matrix) into non-competitive-type IO table (an IO table with a separated import matrix), a useful but sometimes rough “same proportion assumption” is always conducted. Distributing the total import row-wise by using the proportion (intermediate input/row-sum of intermediate inputs) in a competitive-type IO table allows the import matrix to be easily estimated. The assumption is that there is no difference between the distribution share of domestic intermediate goods and imported intermediate goods across sectors. The OECD ICIO tables consist of a time series of international IO data for 62 countries and regions and 34 industries at the basic price. In our exercise, we use the 2005 and 2007 OECD table as our control total for Japan and China, respectively. Again, for simplicity in this exercise, we aggregate the OECD ICIO table from a 62-country, 34-industry table to a table with 5 country groups (China, Japan, U.S., EU, and the rest of the world) and 8 industries. The Chinese regional customs data cover 31 domestic provinces (based on the statistics from more than 400 regional customs offices). The Japanese regional customs data cover 148 customs offices located in different provinces.6 We use the OECD end-use category for commodity classification to rearrange the Harmonized System eight-digit customs import and export data into three categories: intermediate products, consumption products, and capital products. This information will help to more reliably split China 5 http://www.meti.go.jp/english/statistics/tyo/tiikiio/. 6 http://www.customs.go.jp/toukei/info/tsdl.htm. Page 23 of 38 Meng and Yamano Economic Structures (2017) 6:23 and Japan’s regional import and export data by country of origin and destination in the estimation process of the REXICIO tables. For detailed country, region, and sector classifications used in this paper, one can refer to “Appendix 1, Appendix 2.” 3 Results anddiscussion 3.1 China andJapan’s regional value‑added export bycountry/region ofdestination To investigate details of the bilateral structure of China and Japan’s regional export of value added, respectively, we calculate the share of value-added exports of a specific destination by country/region using both gross exports and value-added-based measures. From Fig.3, we can see a significant difference for China’s central region. This result implies that the central region does not directly export a great deal to foreign countries in terms of gross exports, but it can participate in GVCs by providing intermediate products to coastal regions, thereby exporting more value added overseas. Thus, China’s coastal regions have been an important bridge linking GVCs and China’s domestic value chains. Through this linkage, the inland regions can take part in GVCs indirectly by providing support to the coastal regions’ domestic value chains, even though the inland regions may not have an advantage in accessibility to overseas markets when compared with the coastal regions. A similar situation can be observed for Japan’s Tohoku region, as shown in Fig.4. Namely, the degree of Japan Tohoku’s participation in GVCs is much higher than commonly thought since Tohoku’s value added may be indirectly exported to foreign countries through providing parts and components to one of its downstream regions (e.g., Kanto), which further performs processing and exports. 3.2 GVC participation byvarious routes As shown in Fig.2, regional outflows and exports of value added can be further decomposed into four parts according to the route, or segment, of GVCs. Based on calculations using the newly constructed REXICIO table, we show these four parts in Fig.5 for the Chinese case. Clearly China’s regional outflows of value added are mainly achieved through the domestic segment of GVCs (VOD), especially for inland regions. Because of the export-oriented nature of the China Coast region, its domestic segment accounts for just 27% of its total GVCs. Regional outflows of value added created in the international segment of GVCs (VOI) are extremely small for all regions. For example, the value added induced in China’s Guandong province when a household in Shanghai consumes Gross vs. value-added term measure on China’s regional exports (Center, 2007) 0% 20%40% 60%80% 100% China Center's Value-added exports China Center's Gross exports China Northest China West China Coast Japan USA EU ROW Domestic Domestic Foreign Foreign Fig. 3 Gross versus value-added term measure on China’s regional exports (Center 2007) Page 24 of 38 Meng and Yamano Economic Structures (2017) 6:23 final products produced in the USA should be very small. The main reasons are (1) Chinese regional demand for foreign final products is still low; (2) given the large domestic production capacity and the relatively low price of final goods, most final demand can be satisfied through domestic supply; and (3) China’s regional (Guandong) demand for foreign (USA) products can induce foreign (U.S.) production to some extent. However, when foreign countries (USA) produce final goods to satisfy Chinese regional demand, they may not require intermediate goods from other Chinese regions (Shanghai). The feedback effect caused by domestic demand for foreign goods that return to other Chinese regions through the international segment of GVCs is therefore very small. However, when looking at regional value-added exports, we can see that there is not a large difference between the international (VEI) and domestic (VED) segment for all regions in comparison with the difference between VOD and VOI. China’s inland regions seem to export value added mainly through the international segment of GVCs, while China Coast exports value added through the domestic segment of GVCs. This is not Gross vs. value-added term measure on Japan’s regional exports(Tohoku, 2005) 0% 20%40% 60%80% 100% Japan Tohoku's Value-added exports Japan Tohoku's Gross exports Japan Tohoku Japan Kanto Japan Chubu Japan Kinki Japan Chugoku Japan Shikoku Japan Kyushu Japan Okinawa China USA EU ROW Domestic Domestic Foreign Foreign Fig. 4 Gross versus value-added term measure on Japan’s regional exports (Tohoku 2005) Domestic region’s GVC participation by route (China, 2007) 0% 20%40% 60%80% 100% China Northest China West China Center China Coast VODVOIVEI VED Fig. 5 Domestic region’s GVC participation by route (China 2007). VOD: regional value-added outflow through the pure domestic segment of GVCs. VOI: regional value-added outflow through the pure international segment of GVCs. VEI: Regional value-added exports through the pure domestic segment of GVCs. VED: regional value-added exports through the pure international segment of GVCs Page 31 of 38 Meng and Yamano Economic Structures (2017) 6:23 Table 7 Concordance betweenthe OECD ICIO sectors andREXICIO sectors Sector name (OECD ICIO) Code s1 s2 s3 s4 s5 s6 s7 s8 Agriculture Mining and quarrying Life‑related industry Process industry Assembly industry Electricity, gas, andwater supply Construction Other services Agriculture, hunting, forestry, and fishing C01T05AGR √ Mining and quarrying C10T14MIN √ Food products, beverages, and tobacco C15T16FOD √ Textiles, textile products, leather, and footwear C17T19TEX √ Wood and products of wood and cork C20WOD √ Pulp, paper, paper products, printing, and publishing C21T22PAP √ Coke, refined petroleum products, and nuclear fuel C23PET √ Chemicals and chemical products C24CHM √ Rubber and plastics products C25RBP √ Other nonmetallic mineral products C26NMM √ Basic metals C27MET √ Fabricated metal products C28FBM √ Machinery and equipment, nec C29MEQ √ Page 32 of 38 Meng and Yamano Economic Structures (2017) 6:23 Table 7 continued Sector name (OECD ICIO) Code s1 s2 s3 s4 s5 s6 s7 s8 Agriculture Mining and quarrying Life‑related industry Process industry Assembly industry Electricity, gas, andwater supply Construction Other services Computer, electronic, and optical equipment C30.32.33CEQ √ Electrical machinery and apparatus, nec C31ELQ √ Motor vehicles, trailers and semi-trailers C34MTR √ Other transport equipment C35TRQ √ Manufacturing nec; recycling C36T37OTM √ Electricity, gas, and water supply C40T41EGW √ Construction C45CON √ Wholesale and retail trade; repairs C50T52WRT √ Hotels and restaurants C55HTR √ Transport and storage C60T63TRN √ Post and telecommunications C64PTL √ Financial intermediation C65T67FIN √ Real estate activities C70REA √ Renting of machinery and equipment C71RMQ √ Page 33 of 38 Meng and Yamano Economic Structures (2017) 6:23 Table 7 continued Sector name (OECD ICIO) Code s1 s2 s3 s4 s5 s6 s7 s8 Agriculture Mining and quarrying Life‑related industry Process industry Assembly industry Electricity, gas, andwater supply Construction Other services Computer and related activities C72ITS √ R&D and other business activities C73T74BZS √ Public admin. and defense; compulsory social security C75GOV √ Education C80EDU √ Health and social work C85HTH √ Other community, social and personal services C90T93OTS √ Private households with employed persons C95PVH √ Page 34 of 38 Meng and Yamano Economic Structures (2017) 6:23 Appendix 2: China andJapan’s domestic regions See Table8 and Figs.7 and 8. Table 8 Region coverage Regional coverage Description (the numbers in the parenthesis correspond to regional numbers in the maps) China 1. Northeast Liaoning(6), Jilin(7), Heilongjiang(8) 2. West Inner Mongolia(5), Guangxi(20), Chongqing(22), Sichuan(23), Guizhou(24), Yunnan(25), Tibet(26), Shaanxi(27), Gansu(28), Qinghai(29), Ningxia(30), Xinjiang(31) 3. Center Shanxi(4), Anhui(12), Jiangxi(14), Henan(16), Hubei(17), Hunan(18) 4. Coast Beijing(1), Tianjin(2), Hebei(3), Shanghai(9), Jiangsu(10), Zhejiang(11), Fujian(13), Shandong(15), Guangdong(19), Hainan(21) Japan 1. Hokkaido Hokkaido(1) 2. Tohoku Aomori(2), Iwate(3), Miyagi(4), Akita(5), Yamagata(6), Fukushima(7) 3. Kanto Ibaraki(8), Tochigi(9), Gunma(10), Saitama(11), Chiba(12), Tokyo(13), Kanagawa(14), Niigata(15), Yamanashi(19), Nagano(20), Shizuoka(22) 4. Chubu Toyama(16), Ishikawa(17), Gifu(21), Aichi(23), Mie(24) 5. Kinki Fukui(18), Shiga(25), Kyoto(26), Osaka(27), Hyogo(28), Nara(29), Wakayama(30) 6. Chugoku Tottori(31), Shimane(32), Okayama(33), Hiroshima(34), Yamaguchi(35) 7. Shikoku Tokushima(36), Kagawa(37), Ehime(38), Kochi(39) 8. Kyushu Fukuoka(40), Saga(41), Nagasaki(42), Kumamoto(43), Oita(44), Miyazaki(45), Kagoshima(46) 9.Okinawa Okinawa(47) Page 35 of 38 Meng and Yamano Economic Structures (2017) 6:23 Fig. 7 Regions in mainland China Page 36 of 38 Meng and Yamano Economic Structures (2017) 6:23 Appendix 3: Decomposition results ofChina andJapan’s value‑added exports andoutflow See Table9. Fig. 8 Regions in Japan Page 37 of 38 Meng and Yamano Economic Structures (2017) 6:23 Table 9 Decomposition results ofChina andJapan’s value‑added exports andoutflow 2007 Million US$ (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) VOD1 VOD2 VOD3 VOD=(1)+(2)+(3) VOI1 VOI2 VOI=(5)+(6) VO=(4)+(7) VEI1 VEI2 VEI3 VEI=(9)+(10)+(11) VED1 VED2 VED=(13)+(14) VE=(12)+(15) Total=(8)+(16) China Northeast 27,174 46,098 6959 80,231 20 675 695 80,926 38,644 4050 251 42,945 20,885 10,988 74,818 117,763 236,639 Share 11% 19% 3% 34% 0.0% 0.3% 0% 34% 16% 1.7% 0.1% 18% 9% 5% 32% 50% 100% China West 84,528 85,292 8735 178,554 33 996 1029 179,584 61,662 6391 394 68,447 35,758 22,894 127,099 195,546 435,762 Share 19% 20% 2% 41% 0.0% 0.2% 0% 41% 14% 1.5% 0.1% 16% 8% 5% 29% 45% 100% China Center 73,251 95,596 7881 176,728 43 1166 1208 177,937 74,777 7752 471 83,000 37,440 27,420 147,860 230,861 482,367 Share 15% 20% 2% 37% 0.0% 0.2% 0% 37% 16% 1.6% 0.1% 17% 8% 6% 31% 48% 100% China Coast 120,262 89,625 11,168 221,055 239 1491 1730 222,786 266,147 30,066 1825 298,038 292,576 7774 598,388 896,426 1383,628 Share 9% 6% 1% 16% 0.0% 0.1% 0% 16% 19% 2.2% 0.1% 22% 21% 1% 43% 65% 100% 2005Million US$ (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) VOD1 VOD2 VOD3 VOD = (1) + (2) + (3) VOI1 VOI2 VOI = (5) + (6) VO = (4) + (7) VEI1 VEI2 VEI3 VEI = (9) + (10) + (11) VED1 VED2 VED = (13) + (14) VE = (12) + (15) Total = (8) + (16) Japan Hokkaido 15,870 14,875 3612 34,357 68 64 132 34,489 5148 714 9 5871 3315 1311 10,497 16,368 55,873 Share 28% 27% 6% 61% 0.1% 0.1% 0% 62% 9% 1.3% 0.0% 11% 6% 2% 19% 29% 100% Japan Tohoku 39,326 29,136 6559 75,021 206 185 391 75,412 14,761 2236 30 17,026 8669 3318 29,013 46,039 135,821 Share 29% 21% 5% 55% 0.2% 0.1% 0% 56% 11% 1.6% 0.0% 13% 6% 2% 21% 34% 100% Japan Kanto 194,684 142,852 24,644 362,180 966 947 1913 364,093 121,764 18,351 247 140,362 76,273 18,422 235,057 375,419 859,363 Share 23% 17% 3% 42% 0.1% 0.1% 0% 42% 14% 2.1% 0.0% 16.3% 9% 2% 27% 44% 100% Japan Chubu 83,516 61,713 14,238 159,468 676 602 1278 160,746 49,122 7824 105 57,051 32,064 8510 97,626 154,677 363,267 Share 23% 17% 4% 44% 0.2% 0.2% 0% 44% 14% 2.2% 0.0% 16% 9% 2% 27% 43% 100% Japan Kinki 93,733 71,814 15,649 181,196 614 575 1189 182,384 48,980 7372 100 56,451 26,882 9632 92,965 149,416 379,591 Share 25% 19% 4% 48% 0.2% 0.2% 0% 48% 13% 1.9% 0.0% 15% 7% 3% 24% 39% 100% Japan Chugoku 34,390 32,831 8373 75,594 312 301 613 76,207 22,382 3343 46 25,771 10,559 4214 40,543 66,314 164,290 Share 21% 20% 5% 46% 0.2% 0.2% 0% 46% 14% 2.0% 0.0% 16% 6% 3% 25% 40% 100% Japan Shikoku 14,298 14,635 3594 32,527 108 105 213 32,740 7699 1129 15 8844 3680 1524 14,048 22,892 63,119 Share 23% 23% 6% 52% 0.2% 0.2% 0% 52% 12% 1.8% 0.0% 14% 6% 2% 22% 36% 100% Japan Kyushu 35,227 26,421 6443 68,092 332 307 638 68,730 24,552 3690 50 28,291 15,873 2882 47,046 75,337 167,981 Share 21% 16% 4% 41% 0.2% 0.2% 0% 41% 15% 2.2% 0.0% 17% 9% 2% 28% 45% 100% Japan Okinawa 2051 1567 316 3935 12 11 23 3958 895 119 2 1015 931 109 2054 3070 7900 Share 26% 20% 4% 50% 0.1% 0.1% 0% 50% 11% 1.5% 0.0% 13% 12% 1% 26% 39% 100% Page 38 of 38 Meng and Yamano Economic Structures (2017) 6:23 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Received: 20 January 2017 Accepted: 26 July 2017 References Cherubini L, Los B (2013) Regional employment patterns in a globalizing world: a tale of four Italies. Banca d’Italia/University of Groningen, mimeo Dietzenbacher E, Guilhoto J, Imori D (2013) The role of Brazilian regions in the global value chain. Munich Personal RePEc Archive (MPRA Papers), No. 54368 Inomata S, Meng B (2013) Compilation and use of the 2005 transnational interregional input–output tables for China, Japan and Korea. Asian Int Input–output Ser, pp 81–82 Johnson RC, Noguera G (2012) Accounting for intermediates: production sharing and trade in value added. J Int Econ 86(2):224–236 Koopman R, Wang Z, Wei SJ (2008) How much of Chinese exports is really made in China? Assessing domestic value added when processing trade is pervasive. In: NBER working paper, 14109 Koopman R, Wang Z, Wei SJ (2014) Tracing value-added and double counting in gross exports. Am Econ Rev 104(2):459–494 Meng B, Zhang Y, Inomata S (2013a) Compilation and applications of IDE-JETRO’s international input–output tables. Econ Syst Res 25(1):36 Meng B, Wang Z, Koopman R (2013b) How are global value chains fragmented and extended in China’s domestic production networks? IDE Discussion Paper, 424 Meng B, Peters G, Zhi W (2015) Tracing greenhouse gas emissions in global value chains. In: SCID Working paper, No.525; and USITC Working Paper, 2014 12A Miller RE, Blair PD (2009) Input–output analysis: foundations and extensions, 2nd edn. Cambridge University Press, Cambridge Murray J, Wood R (2010) The sustainability practitioner’s guide to input–output analysis. Common Ground Publishing, Champaign OECD-WTO (2013) Measuring trade in value added: an OECD-WTO joint initiative. http://www.oecd.org/industry/ind/ measuringtradeinvalue-addedanoecd-wtojointinitiative.htm OECD, WTO and World Back Group (2014) Global value chains: challenges, opportunities, and implications for policy. In: Report prepared for submission to the G20 Trade Ministers Meeting Pei J, Meng B, Wang F, Xue J (2016) Production Sharing, demand spillovers and CO2 emissions: the case of Chinese regions in GVCs, Singapore. Econ Rev 62(2):10 WTO-IDE (2011) Trade patterns and global value chains in East Asia: from trade in goods to trade in Tasks, printed by the WTO Secretariat Zhang YX, Qi SC (2012) China’s interregional Input–output tables for 2002 and 2007. China Statistics Press, Beijing (in Chinese)