Assessing the Impact of Agri-Global Value Chain Participation on Production Diversification: A Forensic Analysis Sunil Saroj, C. Veeramani, Devesh Roy, Mamata Pradhan, and Abul Kamar ICAE Conference New Delhi, India 6th August 2024 Context Setting  In Global Value Chains (GVCs), products cross international borders to supply foreign markets  Rise of GVCs o manufacturing and services sector o agricultural-food sector  Traditionally, agriculture was treated as a sector producing homogenous products  Agri-food GVC (AGVC) are considered important for outcomes o structural transformation and growth in agricultural productivity o Structural transformation within agriculture comprising diverse crop choices o precursor to larger structural change with shift into non-farm and industrial sectors o Agri-productivity (rapid) - May not be possible to move if crop diversification is postponed Motivation and Hypothesis Motivation and Hypothesis Literature on trade and diversification is thin Role of AGVC has not been done Markets and trade have not been assessed Can GVC be associated with diversification (direction is not known) Empirical ambiguity Need to estimate the direction of production diversification Methodological issues Issues in AGVC – positioning of AGVC Set of activities in AGVC Upstream, downstream and midstream Operations of AGVC Data and Methodology Data • OECD-Trade in Value Added (TIVA) database • covering 1995 to 2020 • Database covers 76 countries, • OECD members, European Union nations, ASEAN countries, and G20 members • Database covers 45 unique sectors, comprising agriculture, machinery, food and beverages, and manufacturing. • Indicators (about 46) from TiVA offer insights into global production networks and supply chains that conventional trade statistics may not capture. Country Names # of Countries Region Austria; Belgium; Belarus, Bulgaria; Croatia; Cyprus; Czech Republic; Denmark; Estonia; Finland; France; Germany; Greece; Hungary; Iceland; Ireland; Italy; Latvia; Lithuania; Luxembourg; Malta; Netherlands; Norway; Poland; Portugal; Romania; Russian Federation; Slovak Republic; Slovenia; Spain; Sweden; Switzerland; Ukraine, and United Kingdom 34Europe Brunei Darussalam; Cambodia; China; Chinese Taipei; Hong Kong; Indonesia; Japan; Korea; Laos PDR; Malaysia; Myanmar; Philippines; Singapore; Thailand; and Viet Nam 15 East and Southeastern Asia Argentina; Brazil; Canada; Chile; Colombia; Costa Rica; Mexico; Peru; and United States 9 North, South, and Central America Cameroon; Côte d'Ivoire; Egypt, Morocco; Nigeria, Senegal; South Africa; and Tunisia; 8Africa Australia; Bangladesh; India; Israel; Jordan; Kazakhstan; Pakistan; New Zealand; Saudi Arabia; and Turkey 10Other Region Sector Classifications for AGVC •Chemical •Electricity •Gas •Steam and AC Supply •Machinery and Equipment •Water Supply •Sewerage and Waste Management Input •Agriculture •Hunting and Forestry •Fishing and Aquaculture Primary •Air Transport •Land Transport •Warehousing •Water Transport Logistics •Food, Beverages and Tobacco Processing Source NameIndicators FAOSTATSimpson Index - Based on Production (Q) FAOSTATSimpson Index - Based on Area (ha) FAOSTATSimpson Index - Based on Production Value ($) FAOSTATAgr Use - Nitrogen (t/ha) FAOSTATAgr Use - Phosphate (t/ha) FAOSTATAgr Use - Potash (t/ha) FAOSTATAgr Use - NPK (t/ha) FAOSTATAgr Use - Insecticides (t/ha) FAOSTATAgr Use - Herbicides (t/ha) FAOSTATAgr Use - Pesticides (t/ha) FAOSTATAgr Use - Pest+Inst+Herb (t/ha) FAOSTAT Agriculture value added per worker (constant 2015 US$) FAOSTATShare of Employment in Agr - Female FAOSTAT Emissions Intensity - AgrFood- Kg CO2eq/kg product FAOSTATEmissions Share - Agrifood Systems - CO2eq FAOSTATTemperature Change in Degree Celsius FAOSTATAgri - Comprehensive Trade Cost FAOSTATAgri - Geometric Avg Tariff FAOSTATAgri - Additional Cost other than Tariff Other Indicators Source NameIndicators Worldwide Governance Index Voice and Accountability Index (%) Worldwide Governance Index Political Stability and Absence of Violence/Terrorism Index (%) Worldwide Governance Index Government Effectiveness Index (%) Worldwide Governance Index Regulatory Quality Index (%) Worldwide Governance Index Rule of Law Index (%) Worldwide Governance Index Control of Corruption Index (%) UNCTADApplied Tariff Faced - Agr Component UNCTADApplied Tariff Faced - Fish Component UNCTAD Applied Tariff Faced - Food & Beverages Component UNCTADApplied Tariff Faced - Fertilizer Component UNCTADApplied Tariff Faced - Machinery Component WDIGDP Per Capita - Current Price USD WDI% of Individuals using internet WDIInflation Rate Tariff & NTMNumber of Products Traded WUI World Uncertainty Index – Reporting Countries (%) WUI Trading with Uncertainty Partner – Based on WUR of Partner Countries (%) Methodology  Panel regression models (FE and RE) are employed to assess the effect of GVC on product diversification o may remain a correlation between the transformed lagged dependent variable and the transformed error term. o Reverse causality like participation in GVC leading to production diversification or production diversification leads to participation in GVC (bit ambiguity). o Dynamic panel model (Maximum Likelihood Estimation)  Alternative methods used for Robustness o first involving standard external instrumental variables (IV). o Often such IVs are either unavailable or are weak. o Use the generalized method of moments (GMM) system estimation o provides reliable estimates using internal instruments  Issue of endogeneity o Employ Lewbel (2012) two-step heteroscedasticity based IV estimator to address these concerns in the panel setup. Specification and estimation model  Lewbel Model o Equation 1 - 𝑌 = 𝛽 + 𝛽 𝑋 + 𝛽 𝑋 … + 𝛽 𝑋 + 𝛾 + 𝛿 + 𝜃 + 𝜇 o Equation 2 (transformed) - 𝑍 = 𝑋 𝑉 + 𝑋 𝑉 + ⋯ + 𝑋 𝑉 o i = country, t = time and s=sector  Identification instrument variable used oLagged variables – Generalized Method of Moments  Fixed effects used oCountry, Time and Sector Measures of AGVC Participation  Hummels et al. (2001) introduced a way to measure participation in GVCs using Input-Output (IO) tables o for both direct value-added trade and indirect value-added trade going through other countries. o susceptible to “double counting”  To address this issue o adopt the GVC participation measure proposed by Koopman et al. (2010), o which considers all sources of value added in total exports and tries to address the problem of “double counting.”  Global Value Chain Participation Index (GVCPI) o considers not only the value added generated directly by the country but also the value added generated in other countries involved in the supply chain. Backward and Forward GVC Participation  GVCPI is defined as: o 𝐺𝑉𝐶𝑃𝐼 = o FVA is value of exports that originates from imported inputs and indicates backward GVC participation o DVX is domestic value added in intermediate goods that are further re-exported by partner country  Backward: use of imported inputs to produce for exports o Example: exports of mobile phones or cars using imported parts  Forward: exports of raw materials and intermediate inputs for further processing and export by other countries o Example: India exports aluminum tubing to Taiwan where it is further used in the production of the bicycle later exported Positioning in GVC  Position of a sector in GVCs (Koopman et al., 2011) o 𝑈𝑝𝑠𝑡𝑟𝑒𝑎𝑚 = 𝑙𝑜𝑔 1 + − 𝑙𝑜𝑔 1 + o +ve sign = upstream and –ve sign = downstream  To examine production diversification o 𝐴𝑔𝑟𝑖𝑓𝑜𝑜𝑑 𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑆𝑖𝑚𝑝𝑠𝑜𝑛 𝐼𝑛𝑑𝑒𝑥 = 1 − ∑ 𝑃 o Other measures of diversification (under process) Findings Heterogeneity Across Sectors BGVC (%) 0.00 10.00 20.00 30.00 40.00 50.00 60.00 Inputs 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 Primary 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00 Logistics 0.00 10.00 20.00 30.00 40.00 50.00 Processed Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam  Heterogeneity across sectors  Developed vs developing countries  Success story o Viet Nam and Thailand  Comparative advantage in BGVC – still gain from the trade. Contd.. FGVC (%) 0.00 20.00 40.00 60.00 80.00 100.00 Inputs Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam 0.00 20.00 40.00 60.00 80.00 100.00 Primary Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam 0.00 20.00 40.00 60.00 80.00 100.00 Logistics Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam 0.00 20.00 40.00 60.00 80.00 100.00 Processed Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam GVC Positioning (log) | upstream; +ve slope or downstream; -ve slope 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Inputs Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 Primary Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Logistics Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 Processed Bangladesh China Germany India Peru Thailand United Kingdom Viet Nam GVC Positioning GVC Participation FGVCBGVCVariables CoefCoefCoefCoef 0.010***0.0040.017***-0.016***Input use (tons/ha) (0.002)(0.003)(0.004)(0.006) -0.008***0.0030.014***0.007 Agriculture value added/worker (lagged) (0.002)(0.004)(0.005)(0.007) 0.022***0.0070.029***-0.042**Share of female in agriculture (0.005)(0.009)(0.010)(0.018) -0.008**-0.005-0.016-0.000Emission share agri-food systems (0.004)(0.010)(0.011)(0.012) 0.002-0.0010.001-0.005Cost to trade (0.002)(0.002)(0.004)(0.008) -0.002-0.000-0.0080.007Tariff (0.002)(0.005)(0.005)(0.009) -0.008***-0.003-0.014***0.007Governance index (0.002)(0.004)(0.005)(0.007) 0.022***0.0070.029***-0.042**GDP per capita (0.005)(0.009)(0.010)(0.018) -0.008**-0.005-0.016-0.000internet penetration (0.004)(0.010)(0.011)(0.012) -0.002-0.000-0.0080.007 WUI — Reporting Country (avg across all countries) (0.002)(0.005)(0.005)(0.009) 21,73621,73621,73621,736Observations 0.3580.2200.2450.226R-squared YYYYCountry, Time and Sector FE Factors Affecting Participation in AGVC  Agriculture inputs o Used in intermediate and processed goods and demand driven by the end consumers like supermarket and food suppliers (upstream)  Agr Value added o increased demand for skilled labor in downstream as well as upstream stages  Trade policy plays a crucial role o Cost to trade o Time to trade  Emission o Environmental footprints primarily lie in purchases Lewbel Method (Y => Product Diversification) 2-Step GMM Estimation (Lagged Instruments) Generated Instruments Models  M4M3M2M1 0.000*0.0000.033***-0.000 Backward GVC (0.000)(0.000)(0.007)(0.000) 0.000**0.0000.098***-0.000 Forward GVC (0.000)(0.000)(0.027)(0.000) 0.000***0.000***0.116***0.000 GVC Participation (0.000)(0.000)(0.035)(0.000) -0.000-0.0010.127***0.005*** GVC Positioning (0.000)(0.000)(0.032)(0.001) Impact of AGVC on Product Diversification  Positive and statistically significant association between product diversification across all measures such as BGVC, FGVC, GVCPI and Positioning.  Brings income growth both backward and forward linkages o Diversify their crop selection, access to inputs and preferences.  Diversify crop selection enhances nutri- sensitive  Change in attributes of demand comprising safety, quality and health  These factors collectively affect the spillovers and the potential for crop diversification. Conclusion and Policy Implications  Phenomenon of GVC are positively affecting the production diversification.  Low-income countries are able to integrate in AGVC o reap benefits of the changing nature of the international trading system.  Viet Nam and Thailand setting a trend and how a small countries is participating in AGVC.  Both FGVC and BGVC are robustly associated with diversified agriculture.  Study can help inform agricultural trade policy in several ways o Towards this trade policy openness on both inputs and outputs i.e., orthodox opening as opposed to heterodox opening is important. o Reduction in cost to trade through digitization can be important for GVC engagement. o Structural transformation process and for a sustainable food system. Thank You