UNDERSTANDING GVCS: INSIGHTS FROM RECENT OECD WORK Javier Lopez Gonzalez, Development Division, OECD Trade and Agriculture Directorate Bangkok 12 th of December 2014
Outline i. How do we capture participation? ii. Why should we care? iii. What are its determinants and what role can policy play? iv. What might the impact of participation be on wage inequality? 2
How do we capture it? 3
Traditionally using trade data Tariff headings labelled parts and components, but: products are not exclusive to one end-use (i.e. think milk or tyres) no indication of how products are combined (linkages between buying and selling sectors) Measured gross and not net (iphone Kraemer et al. (2011)) Nevertheless, trade data still useful. 4
Moving towards value added measures 5
TiVA GVC indicators Indicators of participation: Backward Participation: foreign value added content of exports (buying element of GVCs); and Forward Participation: the domestic value added sold to other countries for these to produce exports (selling element of GVCs). TiVA: current revision: 57 countries, 18 sectors, 5 points in time (1995, 2000, 2005, 2008, 2009) New revision (Dec 2014): + 4 countries, expanded sectors and 2010-2011 data. Other Sources: IDE-JETRO, WIOD, EORA, GTAP 6
Global Matrix of Value Added Trade 7
Participation on the rise 80 70 60 50 40 30 20 10 0 GVC participation 80 70 60 50 40 30 20 10 0 Buying Selling Participation 1995 8
Why should we care? 9
Complementarity between foreign and domestic value added Change in domestic value added share -.04 -.02 0.02.04 services Thailand ligth_manufacturing Chemicals&Fuel Transport_Equipment manufacturing_machinery Electrical_Equipment -.15 -.1 -.05 0.05.1 Change in imported value added share primary Change in domestic value added share -.1 0.1.2.3 ligth_manufacturing servicestransport_equipment manufacturing_machinery Chemicals&Fuel primary Philippines -.1 0.1.2 Change in imported value added share Electrical_Equipment Fitted values Change Dom Fitted values Change Dom Indonesia Change in domestic value added share -.1 -.05 0.05 Chemicals&Fuel ligth_manufacturing Electrical_Equipment Transport_Equipment manufacturing_machinery services primary -.1 -.05 0.05.1 Change in imported value added share Fitted values Change Dom 10
Smaller shares of bigger pies China electrical and optical equipment exports (share) China electrical and optical equipment exports (value) 100% 90% 13% 500.00 450.00 80% 70% 60% 43% 400.00 350.00 300.00 183.69 50% 40% 87% 250.00 200.00 30% 20% 57% 150.00 100.00 247.75 10% 0% 1995 2009 50.00-2.93 19.04 1995 2009 Domestic Foreign Domestic Foreign 11
And links with desirable outcomes VARIABLES (1) (2) (3) log of per capita domestic value added in exports log of export sophistication normalised trade concentratio n indicator Backw ard log of value (lag) 0.0124** (0.00568) backw ard (ratio) 0.192** -0.232*** (0.0914) (0.0488) lnimpyint 9.427** 6.852** -4.032** (4.288) (3.222) (1.720) lnimpyint2-0.502** -0.364** 0.211** (0.224) (0.169) (0.0900) lnimpyintprim 0.0310-0.0663*** -0.00442 (0.0250) (0.0185) (0.00990) lnfdi_inflow 0.000522-0.000723** 0.000125 (0.000458) (0.000342) (0.000183) sh_imp_rta 0.000755-0.0170 0.0190 (0.0351) (0.0261) (0.0140) lnwdi_gdp_capita_constant 0.933*** 0.205*** 0.0542*** (0.0384) (0.0286) (0.0153) lndist_activity -2.221*** -0.354 0.347** (0.355) (0.264) (0.141) Constant -31.63-21.05 15.96* (21.24) (15.94) (8.513) Observations 2,050 2,064 2,064 R-squared 0.814 0.374 0.037 Year FE Y Y Y Reporter FE Y Y Y 12
So what are its determinants and what role can policy play? 13
Preliminaries Cannot just compare participation rates and say that one is more integrated than another need to benchmark by looking at determinants of participation 60 50 Backward Participation 40 30 20 10 0 Buying 14
Structural versus policy factors? Structure and geography: Market size Level of development Degree of industrialisation Distance to main manufacturing hubs in Europe, North America and Asia Trade and FDI policies: Import tariffs, tariffs faced in export markets, engagement in RTAs Openness to inward FDI Other policies: Logistics, IPR protection, infrastructure, institutions, electricity supply, R&D spending 15
Structural factors most important but trade and investment policy also play a role Estimation based on OECD TiVA 16
What about other policies? 17
Role of trade policy within and across developing regions Backward participation 0.6 0.5 Esatern and Southrn Africa (ESA) Middle East and North Africa (MENA) South Asia (SAS) Southeast Asia (SEA) Western and Central Africa (WCA) 0.4 0.3 0.2 0.1 0-0.1 DJI BDI RWA UGA KEN AGO MDG ZMB MUS MOZ ZAF MAR TUN EGY SAU JOR ARE LBN TUR BGD MDV BTN IND NPL PAK LKA VNM KHM MNG CHN PHL THA MYS IDN BRN JPN SGP HKG COG GAB CAF CMR TCD NGA GIN NER MLI SEN TGO BEN CIV BFA Actual backward intergartion Contribution associated with trade policy stance Source: estimations based on EORA 18
What about FDI? Backward participation 0.6 0.5 Esatern and Southrn Africa (ESA) Middle East and North Africa (MENA) South Asia (SAS) Southeast Asia (SEA) Western and Central Africa (WCA) 0.4 0.3 0.2 0.1 0-0.1 RWA BDI MDG KEN MUS DJI UGA ZAF MOZ AGO ZMB SAU TUR ARE EGY MAR TUN JOR LBN NPL BTN IND BGD PAK LKA MDV JPN CHN IDN PHL MNG MYS THA VNM KHM BRN SGP HKG BFA NER SEN GAB BEN TGO CAF MLI CMR GIN NGA CIV TCD COG Actual backward intergartion Contribution associated with investment opennness stance Source: estimations based on EORA 19
And unexplained factors? Backward participation 0.6 0.5 Esatern and Southrn Africa (ESA) Middle East and North Africa (MENA) South Asia (SAS) Southeast Asia (SEA) Western and Central Africa (WCA) 0.4 0.3 0.2 0.1 0-0.1-0.2 ZMB BDI ZAF UGA AGO MDG RWA MOZ MUS KEN DJI EGY LBN MAR TUN ARE TUR SAU JOR PAK LKA BGD NPL IND MDV BTN BRN IDN THA CHN KHM JPN MNG PHL VNM HKG SGP MYS CMR BFA TCD CIV SEN COG GAB BEN MLI GIN NER NGA CAF TGO Actual backward intergartion Under/Over-perfromance Source: estimations based on EORA 20
Can FTAs help? Processed Intermediates goods Vehicles (HS87) 29% 23% 76% Electr. Equip. (HS85) 55% 10% 5% Metals (HS72-83) 38% 11% 10% Textiles (HS50-63) -4% 41% 35% Plastic/Rubber (HS39-40) 33% 16% 28% -20% 0% 20% 40% 60% 80% 100% 120% 140% Reg. Trade Bilateral TA Reg. Trade + TA 21
What might be the impact of participation on wage inequality? 22
The grand trade and inequality debate rages on GVCs have altered the geography of production. Globally, wage inequality has fallen but country experience has been mixed. Old theory (HOS): predicts falling inequality in developing countries but rising inequality in developed countries. New theory (trade in tasks) ambiguous (rise or fall depending on type of linkage) so ultimately an empirical question. 23
Wage inequality has a strong development dimension WIODGINI.1.2.3.4.5.6 IND IDN BRA ROM BGR RUS TUR HUN KOR PRT LTU USA POL SVN ESPAUS CAN NLD BEL CZE GBR AUT FINDNK SWE LUX 7 8 9 10 11 Per Capita GDP (natural logarithm) 95% Confidence Interval Linear Prediction obs 24
Countries with a higher backward participation tend to witness lower wage inequality WIODGINI.1.2.3.4.5.6 BRA IND RUS CHN MEX IDN TUR TWN LVA KOR PRT HUN JPNUSA RO M GRC CYP LTU BGR EST AUS ITA DEU POL CAN ESP SVN FRA NLD GBR A UT CZE BEL I RL SVK FINDNK SWE MLT LUX 0.2.4.6 Backward Participation 95% Confidence Interval Linear Prediction obs 25
but the type of linkage matters Residual 1 -.2 -.1 0.1.2 RUS BRA JUSA PN AUS I DN MEX MLT KOR CYP HUN CHN PRT IND TUR IRL DEU CAN NLD AUT GRC BELSVN ESP EST BGRDNK CZE GBR ITA FRA LVA FI NPOL SVK LTU ROMSWE LUX Residual 2 -.2 -.1 0.1.2 RUS BRA KOR CYP MEX MLT IND CHN P RT USA TUR DEU A UT JPN NLD GRC SVN CAN BEL AUS IDN BGR DNK CZE ESP EST SVK FRA FIN GBR ITA LTUPOL LVA ROM SWE HUN IRL LUX 0.05.1.15.2.25 Low and middle skill backward participation 0.05.1.15 High-Skill Backward Participation 95% CI Fitted values Residuals 95% CI Fitted values Residuals 26
How? Firms that offshore receive a productivity boost that is complementary to the task they offshore. Low-skill task offshoring brings about reductions in wage inequality driven by increases in the labour productivity, and therefore wages, of remaining low-skilled workers. High-skill task offshoring boosts high-skill labour productivity and therefore the wages of remaining high-skill workers resulting in higher wage inequality. Caveat: not capturing unemployment directly but robustness measures using OECD income inequality measures reveal similar effects. 27
Thanks Thanks! Javier.lopezgonzalez@oecd.org This is joint work with: Przemyslaw Kowalski, Alexandros Ragoussis and Cristian Ugarte and Pascal Archard 28
Backward Participation 29
30