Identifying hubs and spokes in global supply chains

Transcription

1 Identifying hubs and spokes in global supply chains with redirected trade in value added Paul Veenendaal Arjan Lejour Hugo Rojas-Romagosa

2 Outline Background and purpose Methodology global input-output analysis focus on trade redirection explanation of trading schemes Results hubs in electronics and other business services hub development over time in China, EU12 and other NAFTA Some conclusions

3 Background Hype in attention for trade in value added (WTO, World Bank, OECD, etc.) Inspiring articles: - J&N: Johnson, R. C. and Noguera, G. (2012), Accounting for Intermediates: Production Sharing and Trade in Value Added, Journal of International Economics, 86(2): KPWW: Koopman, R., Powers,W., Wang, Z. and Wei, S.-J. (2010), Give Credit where Credit is Due: Tracing Value Added in Global Production Chains, NBER Working Paper Predecessor of the current paper is - LRV1: Lejour, A., Rojas-Romagosa, H. and Veenendaal, P. (2012),The Origins of Value in Global Production Chains, study for DG TRADE Today I extend in my presentation the analysis of the paper with - a new concept: factory gate value added exports - redirection via intermediate exports - these extensions are available in a draft paper at the GTAP-website 3

4 Purpose Can we retrieve meaningful information from global input-output tables about the most important hubs and spokes at the level of the industry of end-use? Some examples: - electronics production in China, South-East Asia and East Asia - motor vehicle assembly in EU12, Mexico, Japan, etc. - airplane construction in USA and EU15 We aim to identify at the industry level - the hubs that convert intermediate output imports into exports - their most important suppliers and customers - and the regions that do not supply 4

5 Approach (1) Take the datasets from GTAP for 2001, 2004 and these datasets link national input-output tables with bilateral trade statistics - 57 industries, 84 different countries/regions Construct global input-output tables from the datasets - main proportionality assumption used: all incoming imports at the industry level are allocated to intermediate and final use in proportion to the row of the import matrix of the importer hence, for example, German electronics imports from China and German electronics imports from the US are allocated to use categories in Germany in exactly the same proportions 5 April

6 Approach (2) Derive suitable indicators that identify the hubs and spokes at the industry level - we base these on redirected trade in value added: intermediate value added imports that are exported again by the one but last country in the chain - this country is either the final output producer producing for foreign final users exporting intermediate value added exports via final output exports (type I redirection); focus of the paper or the last exporter to the final output producer producing for the home market exporting intermediate value added imports via intermediate value added exports (type II redirection); extra in my presentation 6 April

7 Structure of the global input output tables For a three region world example with regions: r, and M S S S f f f x S S S f f f x S S S f f f x w r w w x r x x r rr r r r r r r r r r r S is an (industry by industry) matrix of intermediate output deliveries, f a vector of final output deliveries by industry, x a vector of gross outputs by industry and w a row vector of value added by industry 7

8 Leontief arithmetic (condensed notation) We summarize the global table with and define input coefficients for intermediates and value added as Then, using S F x w x A( r, i, s, j) S( r, i, s, j) / x( s, j) and v( r, j) w( r, j) / x( r, j) w as an index for the global aggregate over regions, x Ax f ( I A) f Bf w 1 w w in which B denotes the global Leontief inverse and output use. f w is global final 8

9 Leontief arithmetic: going quadrangular Defining - the exports of final output from to as - the local Leontief inverse of country r as And using - the ^ symbol to denote a matrix with a vector on its main diagonal and zeroes elsewhere - as a toggle which is one if and zero elsewhere rs Consider: gross exports r s f 1 rr (1 )( A B I) fˆ rs rs rs s r s value added exports ˆ 1 rs r rr rs 9

10 With these definitions we obtain: the well-known result that all value added exports are absorbed in final output f 1 ˆ rs f r s because the relation of gross exports of final output with value added exports is so clear, we restrict ourselves to analysing intermediate gross and value added trade (1 ) A B fˆ 1 ˆ rs rs rs s and rs r rr rs 10

11 Then we obtain: the importance of imports in exports Intermediate gross exports Intermediate gross imports We learn four things from this: the VS and VAX-ratio s are complements VS rs sr rs s s s VS sr sr r sr s s s the aggregate trade balance is the same in gross trade and in value added trade: imports in exports cancel they don t cancel in the bilateral trade balance though: hence the imports in Chinese exports to the USA exceed the imports in US exports to China (and this may be just a volume effect) value added imports have passed through the chain but value added exports are just at the beginning of the chain 11

12 Type I redirection in a triangular trading scheme: Redirected imports Absorbed imports Indirect exports to r 12

13 Type I redirection in a quadrangular trading scheme: Redirected imports Absorbed imports Direct exports to Indirect imports of r s Indirect exports via direct trading partner s 13 February

14 Type I redirection in a full trading scheme: Redirected imports Absorbed imports Direct exports to Indirect imports of r s2 s3 sc s Indirect exports through the chain 14

15 Type II redirection in a full trading scheme: Redirected type I imports Absorbed imports Redirected type II imports of 2 3 c s sc 1 Imports from r Imports from 2 Imports from 15

16 Factory gate exports Origin Final user Final producer Last exporter Shipments from 1: Shipments from 2: Shipments from 3: 16

17 Collecting type I and type II redirection from the factory gate exports table Origin Final user Final producer Last exporter Redirected by 1 of which type II excluding exports Redirected by 2 by the first and last Redirected by 3 exporter 17

18 Factory gate exports that are not redirected Origin Final user Final producer First and last exporter

19 Calculating factory gate exports We had to restrict the size of the global system to 12 sectors and 12 regions/countries to make the calculations feasible The decay in the chain was about 6.7 We chose the cut-off at the 4th position in the chain After the cut-off only 0.1% of factory gate exports was still wandering in the chain All figures and tables in this presentation are based on the concept of factory gate exports The figures are very similar to those in the paper 19

20 Composition of global value added exports Own value added in own final output exports Factory gate exports Direct for own final output use Direct for final output exports (type I) Redirected for own final output use (type II) Redirected for final output exports (type I) Total redirection Total

21 Country classification in results EU15: Old member states EU12: New member states OWE: Other western Europe OEE: Russia and other eastern Europe IND: India CHH: China and Hong Kong EAS: East Asia SEA: South-East Asia JPN: Japan USA: United States of America ONA: Other Nafta ROW: Rest of World 22

22 Industry classification in results AGO: ENG: LTM: MLM: CRP: MVH: OTN: OME: ELE: TRA: OBS: OCS: OSR: Agriculture and raw materials Energy Low-tech manufacturing Medium-low tech manufacturing Chemical, rubber and plastic products Motor vehicles and parts Transport equipment nec Machinery and equipment nec Electronic equipment Transport Business services Other commercial services Other services 23

23 Type I and Type II redirection as a % of value added exports,

24 GSFRV Total redirection of value added for electronics end-use, China SEA EAS Average EU12 USA EU15 ONA Japan India ROW OWE OEE SFRV 25

25 GSDRV Redirection of domestic value added for electronics, China EU15 USA ROW SEA EAS Japan Average 5 ONA OEE 0 EU12 India OWE SDRV 26

26 GSFRV Origins of redirected foreign value added for electronics end-use, eu15 eu12 owe oee chh ind eas sea jpn usa ona row Redirector Value added from: eu15 eu12 owe oee chh ind eas sea jpn usa ona row 27

27 GSFRV Final destinations of redirected value added for electronics, eu15 eu12 owe oee chh ind eas sea jpn usa ona row Redirector Value added to: eu15 eu12 owe oee chh ind eas sea jpn usa ona row 28

28 GSFRV Type I redirection of value added for 0BS, EU15 10 EU12 USA OWE China India SEA ROW Average 5 0 ONA EAS OEE Japan SFRV 29

29 GSFRV Type I + type II redirection of value added for OBS, China EU15 USA SEA EU12 ROW OWE EAS Average India Japan ONA OEE SFRV 30

30 China as a redirector, % of globally redirected value added ele ltm ome, 2007 mlm 2004 otn 2001 crp mvh % of intermediate value added imports 31

31 EU12 as a redirector, % of globally redirected value added mvh 10 8 mlm ome 6 4 ltm crp otn ele, % of intermediate value added imports 32

32 ONA as a redirector, % of globally redirected value added mvh ltm mlm crp otn ome, 2007 ele 0 % of intermediate value added imports

33 Factory gate exports: upstream or downstream? % eu15 eu12 owe oee chh ind eas sea jpn usa ona row wld Type II plus direct exports for final use at home Type I 34

34 Value added composition of gross exports % eu15 eu12 owe oee chh ind eas sea jpn usa ona row eu15 eu12 owe oee chh ind eas sea jpn usa ona row 35

35 Some conclusions on the analysis Our analysis yields meaningful results on vertical specialisation at the industry level that are robust over time and seem to make sense We contribute to the literature with - the concept of trade redirection by the last but one country in the chain - the concept of factory gate exports, which seems useful in determining the interrelations in global supply chains (hubs and spokes) characterizing exports as relatively up- or downstream measuring the value added content of trade Type II redirection is still cumbersome to compute Our approach is just to measure what is happening in a meaningful way in the hope that this will be useful in testing the predictions of models that aim to explain what is happening in a meaningful way 36