aa r X i v : . [ ec on . GN ] F e b Research Methods of Assessing Global Value Chains
Centre for Development StudiesTrivandrum, Kerala
This section would use several statistical measures to examine the degree of India’s GVC en-gagement. Those measures can be classiﬁed according to the levels of vertical integration, suchas measures of backward integration, measures of forward integration, and some other statisticalmeasures. I have summarised the main measures of forward and backward dimensions as wellas the other crucial measures of India’s GVC engagement in the following sections (data for allmeasures are easily available for India).
Investigating India’s potential in GVCs, at ﬁrst, we have to use the gross trade ﬂows. It requires tolooking at the top export and import products, classiﬁed at the most disaggregated level. Executionof such initial insights increases the relevance of GVC analysis. In this ﬁrst-cut analysis, we willhandle three issues:1. Compare product-level imports with export values, volumes, and prices of the top tradedproducts (whether exports and imports follow different distributions and the values or vol-umes traded have distinct growth or level) to examine the degree of transformation withindomestic segment of India’s major GVCs (Taglioni & Winkler 2016).2. Use informed classiﬁcations (based on the WITS value chain category, OECD BilateralTrade Database by industry and end-use category, and UN Classiﬁcation of Broad EconomicCategories) to extract as much information as possible from gross trade data (Sturgeon & Memedovi´c1011, Athukorala 2011). Regrouping data in meaningful clusters or categorised by informedclassiﬁcations focusing on consumer goods, capital goods, intermediates, and raw materialswith sector analysis is beneﬁcial (Taglioni & Winkler 2016).3. Document trade ﬂows at the sub-national level to account India’s degree of transformationwithin border when available.These measures do not reveal whether the inputs are used domestically or exported. The followingmeasures address this topic, focusing on imported inputs or foreign value added embodied in grossexports.
This section onward, different measures suggest ways identify the extent to which India and itspeers source – domestically or internationally – the intermediates they use in exporting, which willprovide the ﬁrst indication of their participation in GVCs. The section then shows ways to quantifythe domestic value added embodied in countries’ exports (Baldwin & Lopez-Gonzalez 2015). Keyquestions for sourcing dimension are: Where are India’s exports made, and where is the valuecreated?•
Imported Inputs Embodied in Gross Exports: I2E in Intermediate or Total Imports
Measuring the India’s intermediate imports embodied in its gross exports as a percent-age of the India’s total intermediate imports. Major businesses in GVCs perceive im-ports of goods and services as being important or critical for their exports (OECD-WTO2013).2.
I2E by Source Country
A very useful indicator of GVC participation is the origin of the imported inputs em-bodied in India’s gross exports.3.
Distinguishing between Domestic and Foreign Value Added in Imports
Imported inputs may contain domestic value added that is exported to a foreign loca-tion, processed, and re-imported. Re-importing and re-exporting can be quite importantfor some industries in India.•
Value Added in Gross Exports: Decomposition of Domestic Value Added
Full Decomposition of Foreign Value Added
The breakdown of foreign value added into source countries or industries is useful froma buyer’s perspective because it identiﬁes which foreign sources add the most value toits exports.•
Length of Sourcing Chains:
This measure that reﬂects such multi-country considerations captures by looking at a recur-sive measure of I2E on the sourcing side is the length of value chain sourcing (Fally 2011,De Backer & Miroudot 2013). The TiVA data provide a handy means for comparing theaverage number of production stages in a given industry and country. An increase in GVClength over time suggests that the value chain has become more complex, with stages donein more countries.
Gross import data (UN Comtrade, BACI, WITS), categorized using informed clas-siﬁcations (BEC, parts and components, technical classiﬁcations); International I-O data (WIOD,TiVA, World Bank Export Value Added database); Enterprise surveys or other ﬁrm-level surveys.
Key questions for the selling dimension are: Who are the ultimate customers for India’s valueadded, and to what countries is India exporting its value added? India, for example, exports ironore to China, but part of that product ends up in the United States and Germany rather than China.That is the seller’s perspective.•
Intermediates in Output or Gross Exports:
A ﬁrst basic measure of India’s involvement in the production of inputs, as opposed to ﬁnalgoods, is the share of intermediates in gross output. The share of intermediates in gross ex-ports takes the exporting perspective into account. The measure for India and Indian sectors,and relative to peers can provide a ﬁrst-pass indication of whether India has become a moreimportant supplier in GVCs. 3
I2E Trade in Gross Exports:
The indicator importing to exports (I2E) in gross exports measures intermediates sold bya country to a buyer for use in the buyer’s exports (I2E from the buyer’s perspective) as apercentage of the seller’s gross exports.•
Domestic Value Added in Gross Exports of Third Countries:
It indicates the contribution of domestically produced intermediates to exports in third coun-tries. The only difference is that this indicator accounts only for the seller’s intermediatesthat are domestically produced, whereas in the previous case the intermediates could alsocontain some foreign value added (Gaulier et al. 2013).•
Value Added in Final Domestic Demand:
The Trade in Value Added (TiVA) and World Input-Output databases can also provide anunderstanding of the ﬁnal consumers of India’s value-added activities (OECD 2013).•
Length of Selling Chains:
GVC length on the sales side measure gauges the "upstreamness" of India’s exports roughly,the number of downstream stages between India’s producers and ﬁnal consumers (Antràs et al.2012, Chor 2014).•
Domestic Gap between Buying and Selling Chains
A ﬁnal useful metric is to combine import and export upstreamness to compute the domesticgap between the buying and selling chains of individual sectors.
Production data (national statistics, UN-Stat manufacturing data set, ﬁrm-leveldata); Gross export data (Comtrade, BACI, WITS), categorized using informed classiﬁcations(broad economic category, parts and components, technical classiﬁcations); International I-O data(WIOD, TiVA, World Bank Export of Value Added database) and National I-O data; Enterprisesurveys or other ﬁrm-level surveys. • Additional useful measures of GVC participation:
Beyond the Measures of Buying & Selling Sides in India’s GVCs...1. Illustrating how the buyer and seller dimensions can be combined to quantify an overallindicator i.e. the GVC participation index (Koopman et al. 2010).4. Focusing on network metrics. It shows how a country’s position overall, in a sector, ina speciﬁc GVC, and with respect to individual products can be measured and visualisedusing network metrics (Amador & di Mauro 2015).3. Paying special attention to the role of services in value added (Sáez et al. 2015, Francois et al.2015).4. Measures of direct links in GVCs using ﬁrm-level data – the micro perspective.•
Links in GVCs Using Firm-Level Measures:
1. Multinationals’ Share of Inputs from Domestic Suppliers (Blalock & Gertler 2008,Smarzynska Javorcik 2004, Havranek & Irsova 2011)2. Domestic Suppliers’ Share of Output to Multinationals3. Domestic Suppliers’ Share of Exports4. Domestic Producers’ Share of Imported InputsHowever, the next phase would follow various econometric tests with customised analysis to India-speciﬁc needs and overcoming the limits inherent in the above methodologies.
This section focuses on different determinants of GVC links at the country, sector, and ﬁrm levels.The ﬁrst step decomposes gross export growth into its components. If gross export growth isaccepted as a measure of GVC links on the selling side, the decomposition allows for detecting howmuch of the value added is generated at home and abroad. The second part adopts two measuresof GVC links – GVC integration at the country or sector level and a GVC participation dummy atthe ﬁrm level.
This measure would examine the level of signiﬁcance of gross export growth onto its componentsi.e. direct domestic value added embodied in gross exports (intra-sector), indirect (upstream) do-mestic value added embodied in gross exports, re-imported domestic value added, and foreignvalue added embodied in gross exports. If gross export growth is accepted as a measure of GVClinks on the selling side, the decomposition would allow for detecting how much of the value addedis generated across countries (India and its peer countries) or sectors within India.5ow we should adopt two measures of India’s GVC links – GVC integration at the country orsector level and a GVC participation dummy at the ﬁrm level. This research, indeed, focuses ondifferent determinants of GVC links at the country, sector, and ﬁrm levels.
Here we focus on the following three country characteristics, which, according to the economicliterature, are important determinants of GVC participation: (1) logistics performance, (2) share ofpeople with a tertiary education in the workforce, and (3) geographical distance to the closest globalknowledge centers. However, initial insights can be gathered by assessing the statistical correlationbetween measures of GVC integration with those selected indicators at the country level. Thisanalysis uses the measure of structural integration in GVCs (Amador & di Mauro 2015).1.
GVC Integration and Logistics Performance
Good logistics performance is important because key components of GVC production aretime sensitive, and reliable connectivity allows ﬁrms to connect factories across borders moreefﬁciently. We will use the overall Logistics Performance Index (LPI) to quantify logisticsperformance. The LPI takes into account a country’s customs efﬁciency, quality of trade andtransport infrastructure, ease of arranging shipments, quality of logistics services, ability totrack and trace consignments, and delivery times.2.
GVC Integration and Skill Levels
A skilled workforce is recognised as an important determinant of countries’ success in GVCsbecause it allows producing at the high standards of productivity, efﬁciency, sophistication,and timeliness required to serve global markets. We will use the share of workers withtertiary education to quantify the skill level.3.
GVC Links and Geographical Distance to Knowledge Centers
Countries closer to the hubs in GVCs and to the global centers of knowledge are favoured byeasier access to tacit knowledge. Unlike knowledge embodied in technology, tacit knowledgerequires frequent and continued face-to-face interaction between the staff and managers oflead ﬁrms or turnkey suppliers and those of other ﬁrms in the GVC, and the importance oftacit knowledge increases for more complex tasks. We will use the geographical distancefrom Germany, Japan, and the United States as a proxy for distance from knowledge centers.6 .3 Determinants of Firm-Level GVC Entry
Following the literature on the ﬁrm-level determinants of exporting, the model includes ﬁrm size,ﬁrm age, foreign ownership status, as well as measures of workers’ skills and productivity as deter-minants of GVC participation. Here we also try to amend the Roberts & Tybout (1997) theoreticalmodel on the determinants of exporting.
If ﬁrm-level data – in particular, information on GVC participation – are not available, an alternativecould be to estimate the impact of the policy determinants on sector GVC participation in India.
This section focuses on the role of GVC links for economic upgrading. Three main measures ofeconomic upgrading are adopted: (1) growth of domestic value added embodied in gross exports atthe sector level in the ﬁrst section, (2) level of domestic value added at the sector level in the secondsection, and (3) ﬁrm-level labor productivity in the third section. Different measures of GVClinks are also explored, including GVC measures of structural integration as buyers and sellersin networks, foreign value added embodied in gross exports, domestic value added embodied inexports of third countries, GVC participation index, position in GVCs (upstreamness), domesticlength of sourcing chains, and share of foreign output in a sector.
Do the intensity and nature of GVC links matter for growth in domestic value added that is ex-ported? This question can be explored through econometric analysis from several angles i.e. (a)whether the degree of structural integration in global value-added trade matters, (b) economet-ric analysis can be used to investigate how greater integration of India in GVCs as a buyer – asopposed to weaker integration as a seller (that is, more unbalanced GVC integration) – affectsdomestic value-added growth from gross exports, (c) the analysis can examine more closely therelation between the growth of foreign value added embodied in gross exports and the domesticvalue-added component, (d) it can look at the role of India’s position in the value chain (upstream-ness or distance to ﬁnal demand), and (e) econometrics can be used to investigate the role of thedomestic length of the sourcing chains. 7.
Growth of GVC Participation and Domestic Value Added Embodied in Exports Growth of Balanced GVC Participation and Domestic Value Added Embodied in GrossExports Growth of Foreign and Domestic Value Added Embodied in Gross Exports Growth of Upstreamness and Domestic Value Added Embodied in Gross Exports Growth of Domestic Length of Sourcing Chains and Domestic Value Added Embodiedin Gross Exports
It focuses on the effect of GVC integration, as a buyer and a seller, on domestic value added, alsotaking into account the mediating role of national policy. Domestic value added is generated bycombining labour with capital stock, and is dependent on a country’s technology shifter. The tech-nology shifter is assumed to be a function of international trade and innovation, which is consistentwith the trade literature (Kummritz et al. 2017). Within-Industry Impact of FDI:
The method focuses on the within industry impact of foreign output share on domestic ﬁrmproductivity and the role of mediating factors. Speciﬁcally, Farole & Winkler (2014 b ) ask,what is the potential of global production networks to enhance the productivity of domesticﬁrms?2. Within-Industry Impact of Structural Integration in GVCs:
Similarly, the analysis can be used to examine the effect of GVC participation of an industryon a ﬁrm’s productivity by merging the Farole & Winkler (2014 b ) data set with two sec-tor measures of structural integration, computed by Amador & di Mauro (2015), in India’sGVCs. This section addresses which GVC-oriented industries have a higher demand for labour, such thatintegrating into GVCs in those sectors has a greater potential to create jobs and increase household8ncome (Calì et al. 2016, Calì & Hollweg 2017). We have following measures (using World Bank’sLACEX database) that can be used to identify the impact on labour and wages. The measures arecategorised in two groups: indirect measures of social upgrading, and direct measures of socialupgrading (Taglioni & Winkler 2016).
This subsection presents indirect measurements of the link between GVC participation and labourmarket outcomes. The speciﬁc sectors that are relevant for participation in GVCs can be analysedusing the following methods.1.
Descriptive statistics may be used to assess which sectors are associated with better labourmarket outcomes. I would examine India and its peer countries’ sector averages of the num-ber of employees, wages and salaries, wage rate (wages and salaries divided by the numberof employees), or labour share (wages and salaries as a percentage of value added). Suchstatistics, for example, can be obtained from the United Nations Industrial DevelopmentOrganisation’s Industrial Statistics database.2.
Analysis of Employment-Generating Industries and Their Level of GVC Integration
This analysis may be carried out by running cross-country "controlled correlations" at thesector level, whereby the labour market indicators discussed in the previous section are re-gressed on indicators of GVC involvement while controlling for other factors, such as regionand gross domestic product. The analyst can also run pooled regressions controlling for in-dustry ﬁxed effects to see which industries have more labour-market-enhancing outcomesconditional on GVC involvement.
This subsection presents more direct measurements of the link between GVC participation andlabour market outcomes by drawing on various indicators already developed in the literature (Taglioni & Winkler2016).1.
Labor Content of Gross Exports
This direct measure of social upgrading is the labour content of gross exports. The newlydeveloped World Bank data set on LACEX can be used to explore the social upgrading linked9o GVC participation (Calì et al. 2016). The data set is computed on the basis of the socialaccounting matrix data available in the Global Trade Analysis Project for intermittent yearsbetween 1995 and 2011.2.
Labour Component of Domestic Value Added in Exports
This direct measure of social upgrading, which was developed by the United Nations Confer-ence on Trade and Development (UNCTAD 2013), is the labour cost component of domesticvalue added in exports, which acts as a proxy for the employment generating potential ofexports.3.
Jobs Sustained by Foreign Final Demand
This indicator, jobs sustained by foreign ﬁnal demand, is being developed by OECD-WTOas part of the TiVA database for 40 countries. The indicator calculates the number of jobsin the total economy sustained by foreign ﬁnal demand, which captures the full upstreamimpact of ﬁnal demand in foreign markets on domestic employment. Rather than considerthe domestic value added in total exports (as was the basis of the previous indicator), whichcould be used as intermediates in third countries and be exported as ﬁnal goods, the indicatorconsiders the domestic value added in foreign ﬁnal demand.4.
Jobs Generated by Foreign Trade in GVCs
India’s participation in GVCs can lead to domestic or foreign labour demand. So this indica-tor is the number of jobs generated by India’s trade in GVCs – jobs generated domesticallyand abroad – using the World Input-Output Database (WIOD). The sources of employmentcreation from international trade are labour demand from ﬁnal goods trade and trade in in-termediates, or the result of India’s GVC participation (Jiang & Milberg 2013).5.
Jobs in GVC Manufacturing
The jobs in the GVC manufacturing indicator will present a broader picture of the structureof employment in GVCs within India and its peers using WIOD. It is the most direct measurein the literature of the domestic employment impacts of manufacturing GVC participationThe indicator would measure – directly and indirectly – the number of GVC jobs involved inthe production of ﬁnal manufacturing goods (also known as manufactures), as well as theirsector of employment in India (Timmer et al. 2014).However, the quality of the assessment depends on the methodology that is applied, which, in turn,depends heavily on data availability. 10
Policy Implications of GVCs
Engagement in GVCs would not bring economic prosperity in an automated manner. It requiresmuch more value added from India’s potential productive factors and upgrading quality & quantityof those factors with a strong distributional aspect of socioeconomic opportunities and outcomes.These challenges truly create the scope for policy discussion. However, the policy options needa strategic framework to maximise the gains from GVC engagement. To develop an effective andsustainable strategy of GVC participation, we must identify key binding constraints and suggestthe necessary policy and regulatory interventions as well as infrastructure and capacity building.That’s why this part of my thesis points out three distinct focus areas with corresponding objectivesand challenges (Taglioni & Winkler 2016).
This ﬁrst focus discusses ways for India and its peers to enter global production networks. Thoseavenues include ways to attract foreign investors, as well as strategies to enhance the participationof domestic ﬁrms in GVCs. Suggestions for entering GVCs encompass measures to ensure thatIndia can offer world-class links to the global economy and create a friendly business climate forforeign tangible and intangible assets.1. Creating World-Class GVC Links – jump-starting GVC entry through the creation of EPZsand other competitive spaces (Milberg & Winkler 2013), attracting the “right” foreign in-vestors as foreign investors vary in their potential to deliver spillovers (Farole & Winkler2014 a ), helping domestic ﬁrms ﬁnd the “right” trade partners and technology abroad, im-proving connectivity to international markets (OECD-WTO 2013).2. Creating a World-Class Climate for Firms’ Assets – ensuring cost competitiveness whileavoiding the trap of low-cost tasks, improving drivers of investment and protecting foreignassets, organising domestic value chains and improving the quality of infrastructure andservices (Cattaneo et al. 2013). This second focus discusses ways for India and its comparator countries to lever their positionin GVCs to achieve higher value addition through economic upgrading and densiﬁcation. Theconcept of economic upgrading is largely about gaining competitiveness in higher-value-addedproducts, tasks, and sectors. Densiﬁcation involves engaging more local actors (ﬁrms and workers)11n the GVC network. Strengthening GVC–local economy links, absorptive capacity, and skillscontributes to the overall goal to increase India’s value added that results from GVC participation.1. Strengthening GVC–Local Economy Links on the Buyer’s and Seller’s Sides (Farole & Winkler2014 a ).2. Strengthening Absorptive Capacity – maximizing the absorption potential of local actors tobeneﬁt from GVC spillovers (Farole & Winkler 2014 a ), Fostering Innovation and BuildingCapacity (Cattaneo et al. 2013).3. Creating a world-class Workforce through skill development (Cattaneo et al. 2013). The third one focuses on social and environmental sustainability of India’s GVCs. Labour mar-ket–enhancing outcomes for workers at home and more equitable distribution of opportunities andoutcomes create social support for a reform agenda aimed at strengthening India’s GVC partici-pation. Climate-smart policy prescriptions and infrastructure can mitigate the challenges for ﬁrmsfrom climatic disruptions, ensuring the long-term predictability, reliability, and time-sensitive de-livery of goods necessary to participate in GVCs.• Creating a World-Class Workforce – promoting skill development, social upgrading, andequitable distribution of opportunities and outcomes.• Implementing Climate-Smart Policies and Infrastructure through proper environmental reg-ulation and innovation in green technologies.
Now I have to provide a summary of the probable methodologies available to carry out the GVCs’assessment and their content in my thesis.•
Value added by broad sector, employment by broad sector, laborproductivity by sector, FDI, exports and imports (% of GDP), exports and imports by broadeconomic category, and other informed classiﬁcations.•
Export market share growth, push, and pull factors:
Export market share growth; decom-position in push and pull factors using shift-share methodologies.12
Measuring Competitiveness in GVCs:
Trade in main GVCs, exports of GVC products rel-evant to India, top ﬁve exports, including country dimensions and follow-up analysis ofinteresting patterns (such as product-speciﬁc analysis).•
Worldwide trade network, country trade network for sector of interest(main buyers), country trade network for sector of interest (main suppliers), and Extensionto more sectors (four or ﬁve, maximum).•
Trade in value-added indicators:
Domestic value added in gross exports (total growth andby sector), decomposition, foreign value added in gross exports, domestic value added inthird countries’ exports, sourcing and selling patterns, value added by destination, importand export upstreamness and gap, contribution of direct and indirect domestic value addedand foreign value added to gross export growth.•
Impact of structural integration in GVCs (network measure) ondomestic value added embodied in exports and gross exports. Probabilistic model of entryin GVCs. Impact of GVC integration (foreign value added in gross exports, domestic valueadded in third countries’ exports) on value added and the role of national policies. Impact ofGVC integration (imported input share, export share, etc.) on labor productivity and the roleof absorptive capacity.•
Role of services in GVCs:
Zoom into the services dimension of GVC analysis•
Product-speciﬁc case study:
Value chain mapping and country positioning, historical/currenttrends, stakeholder/actor analysis, challenges and opportunities, future implications, policyimplications.•
Policy section: policy suggestions based on GVC analysis, screening of policy performanceindicators etc.Indeed, I need a sound knowledge base in policy and strategic issues related to trade in general andGVCs in particular, as well as technical skills in analyzing trade and production data at the macroand ﬁrm levels. Ideally, I should have some knowledge in GVC analysis and, most important,in depth country knowledge. If I intend to do in-depth technical analysis as well, technical toolsmust be learned. If I consider certain methodologies essential from the outset, then bringing inspecialized technical expertise to lead those components may be useful.
Amador, J. & di Mauro, F. (2015), ‘The age of global value chains: maps and policy issues’,
VoxEUeBook, Centre for Economic Policy Research .13ntràs, P., Chor, D., Fally, T. & Hillberry, R. (2012), ‘Measuring the upstreamness of productionand trade ﬂows’,
The American Economic Review (3), 412–416.Athukorala, P.-c. (2011), ‘Production networks and trade patterns in east asia: Regionalization orglobalization?’,
Asian Economic Papers (1), 65–95.Baldwin, R. & Lopez-Gonzalez, J. (2015), ‘Supply-chain trade: A portrait of global patterns andseveral testable hypotheses’, The World Economy (11), 1682–1721.Blalock, G. & Gertler, P. J. (2008), ‘Welfare gains from foreign direct investment through technol-ogy transfer to local suppliers’, Journal of International Economics (2), 402–421.Calì, M., Francois, J., Hollweg, C. H., Manchin, M., Oberdabernig, D. A., Rojas-Romagosa, H.,Rubinova, S. & Tomberger, P. (2016), ‘The labor content of exports database’, World Bank PolicyResearch Working Paper 7615 .Calì, M. & Hollweg, C. H. (2017), ‘How much labor do south african exports contain?’.Cattaneo, O., Gerefﬁ, G., Miroudot, S. & Taglioni, D. (2013), ‘Joining, upgrading and being com-petitive in global value chains: a strategic framework’,
World Bank Policy Research WorkingPaper 6406 pp. 1–50.Chor, D. (2014), ‘Where are countries positioned along global production lines?’,
MacroeconomicReview pp. 94–99.De Backer, K. & Miroudot, S. (2013), ‘Mapping global value chains’.Fally, T. (2011), ‘On the fragmentation of production in the us’,
University of Colorado mimeo .Farole, T. & Winkler, D. (2014 a ), Making foreign direct investment work for Sub-Saharan Africa:Local spillovers and competitiveness in global value chains , World Bank Publications.Farole, T. & Winkler, D. (2014 b ), ‘The role of mediating factors for fdi spillovers in developingcountries: Evidence from a global dataset’.Francois, J., Manchin, M. & Tomberger, P. (2015), ‘Services linkages and the value added contentof trade’, The World Economy (11), 1631–1649.Gaulier, G., Santoni, G., Taglioni, D. & Zignago, S. (2013), ‘In the wake of the global crisis:evidence from a new quarterly database of export competitiveness’.Havranek, T. & Irsova, Z. (2011), ‘Estimating vertical spillovers from fdi: Why results vary andwhat the true effect is’, Journal of International Economics (2), 234–244.Jiang, X. & Milberg, W. (2013), ‘Capturing the jobs from globalization: trade and employment inglobal value chains’. 14oopman, R., Powers, W., Wang, Z. & Wei, S.-J. (2010), Give credit where credit is due: Trac-ing value added in global production chains, Technical report, National Bureau of EconomicResearch.Kummritz, V., Taglioni, D. & Winkler, D. (2017), ‘Economic upgrading through global value chainparticipation’.Milberg, W. & Winkler, D. (2013), Outsourcing economics: global value chains in capitalist de-velopment , Cambridge University Press.OECD (2013), ‘Science, technology and industry scoreboard 2013’.OECD-WTO (2013), ‘Aid for trade at a glance 2013’.Roberts, M. J. & Tybout, J. R. (1997), ‘The decision to export in colombia: an empirical model ofentry with sunk costs’,
The American Economic Review pp. 545–564.Sáez, S., Taglioni, D., Van der Marel, E., Hollweg, C. H. & Zavacka, V. (2015),
Valuing servicesin trade: a toolkit for competitiveness diagnostics , World Bank Publications.Smarzynska Javorcik, B. (2004), ‘Does foreign direct investment increase the productivity of do-mestic ﬁrms? in search of spillovers through backward linkages’,
The American Economic Re-view (3), 605–627.Sturgeon, T. J. & Memedovi´c, O. (2011), Mapping global value chains: Intermediate goods tradeand structural change in the world economy , United Nations Industrial Development Organiza-tion.Taglioni, D. & Winkler, D. (2016),
Making Global Value Chains Work for Development , Vol. Tradeand Development series, Washington, DC: World Bank.Timmer, M. P., Erumban, A. A., Los, B., Stehrer, R. & de Vries, G. J. (2014), ‘Slicing up globalvalue chains’,