## wpiea2023073-print-pdf

## Source details

**Canonical URL:** [wpiea2023073-print-pdf](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023073-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023073-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023073-print-pdf.pdf.json)

---

### Introduction and motivation
- Geoeconomic context:
  - Globalization stalled after the Global Financial Crisis (GFC) and may be reversing due to military conflicts, skepticism about globalization, populism, and protectionism (examples: Brexit, trade war between the United States and China).
  - COVID-19 and the war in Ukraine increased geopolitical splits and uncertainty about globalization.
  - Term introduced: “geoeconomic fragmentation” (GEF) — policy-driven reversal of global economic integration guided by strategic considerations.
- Purpose:
  - Quantify economic costs of fragmentation from an international trade perspective focusing on production and trade of commodities using a novel multi-country multi-sector model with input-output linkages and a new granular commodity dataset.

### Data construction and features
- Coverage and structure:
  - Updated data to 2019 (most recent pre-pandemic year) from 2016 in Fally and Sayre (2019, FS).
  - Reconciled with an aggregate sector in the Eora26 database.
  - Final dataset: 136 primary commodities + 24 manufacturing and service sectors for 145 countries.
  - Commodity disaggregation: 54 mining commodities; 82 agricultural commodities.
- Sources and processing:
  - MRIO starting point: Eora26 multi-region input-output table (26 sectors, 190 countries).
  - Trade data: BACI database (200 countries, HS 6-digit) — values (USD) and quantities (metric tons).
  - Agricultural production: FAO statistics for 173 products across 245 countries.
  - Mining production: World Mineral Production dataset (British Geological Survey).
  - Natural gas trade: IEA.
  - Mining commodity valuation steps: convert quantities to metric tons; derive exporter-commodity unit values from BACI; replace observations > two standard deviations above the mean with the median; Values = quantities × unit values.
- Consistency rules:
  - If sum of trade values < Eora aggregate: allow residual difference as an “other” category.
  - If sum of trade values > Eora figure: scale down exports or self-trade proportionately.
  - If export value > production value: set self-trade = 0.
- Final dataset features (2019):
  - Values (USD) at exporter sector/commodity − importer level including self-trade.
  - 24 aggregated manufacturing and services sectors; 82 agricultural commodities; 54 mining commodities; 145 countries.
  - Crosswalks: hand-developed mapping from production and HS codes to commodity list.
- Definition:
  - Self-trade: value of production consumed in the producing country (i.e., not exported).

### Model structure and methodological choices
- Model type and production structure:
  - Quantitative multi-country multi-sector trade model distinguishing commodities and non-commodities.
  - Non-commodities use labor, commodities, and non-commodity intermediate inputs; non-commodities are consumed as final goods.
  - Commodities rely on labor and non-commodity intermediate inputs.
- Key modelling departures and assumptions:
  - Incorporates sectoral input-output linkages to capture feedback loops and amplify transmission of shocks.
  - Uses proportionality assumptions in use of intermediate inputs (following Cunat and Zymek (2022)).
  - Low elasticity of substitution between commodities and other inputs in non-commodity production to generate low price elasticity of demand for commodities.
  - Static model using exact hat algebra; no endogenous capital accumulation or technological change; initial aggregate trade imbalances are allowed but fixed.
- Analytical first-order insights:
  - Fragmentation has larger impact on prices in countries/sectors that lose access to a larger share of pre-fragmentation supply.
  - Impact is larger where the elasticity of substitution between foreign and domestic products (the ‘trade elasticity’) is lower.
  - Country real GDP effect decomposes into: (i) direct effect of import prices on final goods, (ii) amplification through input-output linkages, and (iii) effect on prices of commodities.
  - First-order change in local price index from elimination of trade between two blocs in sector k:
    - d ln P_n^k ≈ (1 / θ_k) ln(1 / (1 − π_n,otherbloc^k))
    - π_n,otherbloc^k = share of initial expenditures of country n on goods from countries in the other bloc.

### Calibration and elasticities
- Trade elasticities:
  - Baseline long-run elasticities: sector-level estimates from Fontagné et al. (2022).
    - Mining commodities: -3.41
    - Agricultural commodities: -2.91
    - Maximum across sectors reported: -10.56
  - Short-run vs. long-run:
    - Boehm et al. (forthcoming) estimate elasticities can take 7 to 10 years to stabilize; elasticities after one year equal 36 percent of the long-run value.
    - Authors set short-run trade elasticity = 36 percent of long-run value (conservative relative to Andersen and Yotov (2023) which estimate ≈ 10 percent).
  - Long-run trade elasticity mean used in robustness: 6.00; literature bounds for robustness: maximum 6.6 (Broda and Weinstein, 2006) and minimum 2.12 (Boehm et al., forthcoming).
- Demand elasticities:
  - CES elasticity of substitution between inputs set equal to 0.2 (conservative side of the mode surveyed by FS).
  - Approximation relating to price elasticity of demand:
    - ∂ ln(d_mnkj / p_mnkj) / ∂ ln p_mnkj ≈ η (with η = 0.2 by calibration).

### Scenario design and modelling assumptions
- Sector definitions and barriers:
  - High-tech sectors: defined using OECD (2011) classification; highlighted sectors: electronics and machinery, and transport equipment.
  - Counterfactuals introduce infinite trade costs for targeted sectors to ensure zero trade between relevant countries/groups.
- Illustrative scenarios:
  - Scenario A: Trade barriers on energy sector.
  - Scenario B: Trade barriers on high-tech sectors.
  - Scenario C: Trade barriers on agriculture goods.
  - Strategic Decoupling (SD):
    - USA and EU impose barriers to prohibit trade of all goods and services from Russia and vice versa.
    - Additional barriers on trade in high-tech sectors between the US-EU and China.
    - No changes in trade relations between US, EU, China, Russia, and RoW.
  - Geo-economic Fragmentation (GEF):
    - USA and EU impose barriers on all trade with China.
    - RoW forced to join one of the two groups (US-EU or Russia-China) based on historical trade intensities; tariffs on all sectors by the other group.
- Short-run vs. long-run timing:
  - Short-run corresponds to one year after barrier increase with short-run trade elasticity set to 36 percent of long-run for strategic sectors.
  - Long-run corresponds to 10 years after barrier increase with full long-run trade elasticities.

### Main quantitative results and key statistics
- Importance of granularity:
  - Welfare loss more than doubles for LICs when incorporating disaggregated commodity trade versus standard IO.
  - Relative increases in welfare losses from aggregated to disaggregated commodity treatment:
    - AEs: 4 percent
    - EMs: 25 percent
- Global and group-level outcomes (selected exact figures):
  - Global output losses across fragmentation scenarios: between 0.3 percent and 2.3 percent in the long run depending on scenario.
  - Comparison baseline vs. global autarky:
    - Output losses more than double for LICs.
    - Welfare losses increase by 4 percent for Advanced Economies (AEs).
    - Welfare losses increase by 25 percent for Emerging Market economies (EMs).
  - Sensitivity to trade elasticity calibration:
    - Geo-fragmentation global GDP loss range: 1.9 percent to 7.0 percent (alternative literature estimates), baseline estimate 2.3 percent.
- Scenario-specific outcomes (preserving exact figures from source):
  - Scenario A (energy trade eliminated):
    - Short-run output decreases: EMs: 10.1 percent; AEs: 7.8 percent; LICs: 6.3 percent.
    - Long-run impacts range from 1.1 to 2.1 percent across country groups.
  - Scenario B (zero high-tech trade):
    - AEs experience the largest output loss (reflects broader representation of high-tech trade for AEs).
  - Scenario C (zero agricultural trade):
    - Emphasizes importance of international trade in agriculture especially for LICs.
  - Strategic Decoupling (SD):
    - Short-run output reductions: AEs: 0.7 percent; EMs: 0.8 percent.
    - Long-run output reduction: 0.3 percent for AEs and EMs.
    - LICs benefit slightly over both time periods due to trade diversion.
    - Global output: fall of 0.8 percent in the short run and 0.3 percent in the long run (baseline elasticities).
  - Geo-economic Fragmentation (GEF):
    - Aggregate long-run reduction in real GDP: 2.3 percent.
    - Aggregate short-run reduction in real GDP: 4.8 percent.
    - Short-run losses: AEs: 4.2 percent; EMs: 5.2 percent; LICs: 10.8 percent.
    - Long-run losses: AEs: 2.1 percent; EMs: 2.5 percent; LICs: 4.3 percent.
- Regional outcomes (baseline, long-run elasticities):
  - SD yields positive impacts via trade diversion in several regions; largest positive: Middle East and Central Asia; largest negative: Emerging Market Europe (losses reduce from 3.8 percent of GDP to 3.0 percent if Russia and Turkey are excluded).
  - GEF produces negative impacts across all regions; Latin America and the Caribbean show the smallest negative impact; Emerging Market Europe shows the largest losses.

### Robustness checks and sensitivity
- Trade elasticity robustness:
  - Long-run trade elasticity mean used: 6.00.
  - Literature bounds used: maximum 6.6; minimum 2.12.
  - GEF welfare-cost range: 1.9 percent (high elasticity case) to 7.0 percent (low elasticity case), baseline 2.3 percent.
- Inter-bloc aggregate imbalances:
  - Re-running baseline with scaled imbalances so surpluses and deficits cancel within each bloc yields results very close to baseline (scaling factors on the order of 15 percent).
  - Table (GEF, long-run elasticities) — preserved values:
    - AE: Baseline -2.14; No inter-bloc imbalances -2.09; Geopolitical blocs -3.40
    - EM: Baseline -2.48; No inter-bloc imbalances -2.36; Geopolitical blocs -3.03
    - LIC: Baseline -4.25; No inter-bloc imbalances -4.12; Geopolitical blocs -4.17
    - Global: Baseline -2.34; No inter-bloc imbalances -2.25; Geopolitical blocs -3.20
- Alternative bloc formation:
  - Assigning blocs via Ideal Point Distance (IPD, geopolitical closeness) increases global aggregate losses to 3.2 percent of global GDP (higher than baseline bloc formation based on trade intensities).

### Conclusions and policy implications
- Main conclusions:
  - Not accounting for granular commodity production and trade linkages leads to underestimation of output losses from trade fragmentation.
  - LICs face losses more than twice those estimated with aggregate commodity sectors; AEs and EMs face costs that are 4 percent and 25 percent larger respectively under granular treatment.
  - Output losses increase with the depth of fragmentation; LICs are disproportionately affected, especially under severe fragmentation forcing exclusive bloc alignment.
  - Short-run adjustment costs can be large due to lower short-run elasticities (example short-run LIC cost of 10.8 percent under GEF).
  - Baseline loss estimates are conservative relative to ranges in the literature; elasticity assumptions materially affect outcomes.
  - Bloc formation based on geopolitical allegiances raises global losses; closing inter-bloc trade imbalances does not materially change results.
- Recommended future research directions:
  - Microfound short- and long-run elasticity differences via quadratic adjustment costs in a dynamic system.
  - Introduce endogenous capital formation and technological progress to capture long-run welfare effects from short-run adjustments.
  - Relax exogenous overall trade imbalance assumption.
  - Explore endogenous bloc formation algorithms where countries iteratively choose blocs based on evolving welfare benefits.

*Source: wpiea2023073-print-pdf — IMF Working Paper excerpt.*

### References .............................................................................................................

### References

### Introduction and motivation
- Globalization stalled after the Global Financial Crisis (GFC) and may be reversing due to military conflicts, skepticism about globalization, populism, and protectionism (examples given: Brexit, trade war between the United States and China).
- COVID-19 and the war in Ukraine increased geopolitical splits and uncertainty about globalization.
- Aiyar and others (2023) introduce the term “geoeconomic fragmentation” (GEF) to describe policy-driven reversal of global economic integration guided by strategic considerations.
- Purpose of the paper: quantify economic costs of fragmentation from an international trade perspective, with a focus on production and trade of commodities using a novel multi-country multi-sector model with input-output linkages and a new granular commodity dataset.

### Data construction and features
- Dataset extensions and features:
  - Updated data to 2019 (most recent pre-pandemic year) from 2016 in Fally and Sayre (2019, FS).
  - Modified commodity list to reflect the most “upstream” products to capture commodity characteristics (e.g., geographical concentration).
  - Reconciled dataset with an aggregate sector in the Eora26 database (Lenzen et al., 2012, 2013).
  - Coverage: 136 primary commodities + 24 manufacturing and service sectors for 145 countries.
- Motivation: granular commodity detail is essential to capture spillovers and concentration effects in production and trade.

### Model structure and methodological choices
- Model type: quantitative multi-country multi-sector trade model, distinguishing commodities and non-commodities.
- Production structure:
  - Non-commodities use labor, commodities, and non-commodity intermediate inputs; non-commodities are consumed as final goods.
  - Commodities rely on labor and non-commodity intermediate inputs.
- Key modelling departures and assumptions:
  - Incorporates sectoral input-output linkages to capture feedback loops in production and amplify transmission of shocks.
  - Uses proportionality assumptions in use of intermediate inputs following Cunat and Zymek (2022) to reduce data burden.
  - Introduces low elasticity of substitution between commodities and other inputs in non-commodity production to generate low price elasticity of demand for commodities.
- Analytical insights:
  - Up to a first-order approximation, fragmentation has a larger impact on prices in countries/sectors that lose access to a larger share of pre-fragmentation supply.
  - Impact is particularly large where elasticity of substitution between foreign and domestic products (the ‘trade elasticity’) is lower.
  - Country real GDP effect decomposes into: (i) direct effect of import prices on final goods, (ii) amplification through input-output linkages, and (iii) effect on prices of commodities.

### Calibration and elasticities
- Demand elasticities for commodity sectors sourced from FS, which conducts a meta-analysis.
- Trade elasticities calibrated based on Fontagné et al. (2022), described as conservative and in line with other literature estimates.
- Sector-specific trade elasticity values noted:
  - Mining and Quarrying: 3.4
  - Agriculture: 2.9
  - Average across sectors: 6
- Consideration of short-run trade elasticities (lower than long-run due to adjustment costs), following Boehm et al. (forthcoming).

### Key quantitative findings and comparative equilibria
- Accounting for granular commodity trade and production substantially increases adverse impact of trade fragmentation versus models that assume perfect substitutability among commodities.
- Comparing baseline (2019 global trade barriers) to global autarky:
  - Output losses more than double for Low-Income Countries (LICs).
  - Welfare losses increase by 4 percent for Advanced Economies (AEs).
  - Welfare losses increase by 25 percent for Emerging Market economies (EMs).
- Fragmentation scenarios summarized:
  - Mild fragmentation (“Strategic Decoupling”):
    - No trade between US-EU and Russia.
    - No trade in high-tech sectors between US-EU and China.
    - Rest of World (RoW) can trade with both groups.
  - Severe fragmentation (“Geo-economic Fragmentation”):
    - No trade between EU-US and Russia-China.
    - RoW joins one of the two groups depending on trade link strength, resulting in zero trade with the other group.
- Global output loss estimates:
  - Global output losses fall by between 0.3 percent and 2.3 percent in the long run depending on fragmentation scenario.
  - LICs: in the mild scenario LICs benefit from trade diversion; in the severe scenario LICs’ output drops by 4.3 percent over the long run.
  - Sensitivity to trade elasticity calibration: global GDP loss in the more severe scenario ranges between 1.9 percent and 7.0 percent when using alternative literature estimates of trade elasticities.
- Sectoral vulnerability:
  - Advanced Economies (AEs) and Emerging Market Economies (EMEs) most vulnerable to disruption in trade in energy and high-tech manufacturing sectors.
  - LICs most vulnerable to barriers to trade in agricultural goods.

### Comparison with related literature and extensions
- IMF (2022): eliminating trade in aggregated high-tech and energy sectors across rival blocs (based on UNGA vote) yields about 1.2 percent world GDP loss, increasing to 1.5 percent when barriers extended to other sectors.
- Cerdeiro et al. (2021): welfare costs from three fragmentation layers range from zero to 8.5 percent when accounting for all layers.
- Bekkers and Goes (2022): range of losses from 0.4 percent to 12 percent depending on scenario and country.
- Javorcik et al. (2022): ‘friend-shoring’ yields output losses between 0.1 percent and 4.6 percent depending on country and scenario.
- Sanctions and energy-specific studies:
  - Bachmann et al. (2022): 30 percent gas supply shock (Russian gas shut-off) affects German Gross National Expenditure by -0.7 to -2.3 percent depending on elasticity of substitution.
  - Lan et al. (2022): Germany’s real GDP falls by 1.4 percent in 2022 and 2.7 percent in 2023 under limited household consumption adjustment.
  - DiBella et al. (2022): multi-sector partial equilibrium model finds average effect of -1.8 percent on EU GDP.
  - Albrizio et al. (2022): substituting Russian gas with LNG reduces adverse impacts by a factor of five but with significant global spillovers.
- Distinction of this paper: focuses on international trade channel with granular commodity treatment (rather than treating commodities as perfect substitutes), calibrates with recent trade elasticities literature, and captures short- and long-run costs.

### Structure of the paper (as outlined)
- Section 2: dataset creation and calibration of elasticities.
- Section 3: model.
- Section 4: fragmentation scenarios.
- Section 5: results and robustness checks.
- Section 6: conclusion.

*IMF WORKING PAPERS — Fragmentation in Global Trade (excerpt).*

### 2.  Data

### 2.  Data

### Production and trade data
- Data construction purpose: produce a new dataset covering 2019 (the latest year prior to the start of the COVID pandemic) with bilateral trade data (including self-trade) for all sectors and detailed disaggregation of commodity sectors.
- Starting MRIO: Eora26 multi-region input-output table (MRIO) with 26 sectors and 190 countries.
- Commodity disaggregation objective: break apart Eora26’s ‘Mining and Quarrying’ and ‘Agriculture’ sectors into detailed individual commodities.
- Commodity list: based on that used by FS with modifications to use the most upstream commodities possible; final counts:
  - 54 mining commodities.
  - 82 agricultural commodities.
  - Full list provided in Appendix I (in source).
- Trade data:
  - BACI database produced by CEP II, bilateral trade flows for 200 countries at the harmonized system (HS) 6-digit level.
  - BACI provides values (USD) and quantities (metric tons).
- Production data:
  - Agricultural: FAO statistics with production value data for 173 (raw and processed) agricultural products across 245 countries.
  - Mining: World Mineral Production dataset from British Geological Survey (BGS) with production quantities for more than 70 mineral commodities by country.
  - Natural gas trade data: taken from IEA for measurement challenges with pipelines and financial transactions.
- Conversion and valuation steps for mining commodities:
  - Convert all quantity units into metric tons using standard conversion ratios.
  - Derive exporter-commodity specific prices using unit values from BACI (unit value = total value of all mapped 6-digit HS codes divided by total quantity when multiple HS codes map to a single commodity).
  - Replace observations greater than two standard deviations above the mean with the median value to bound prices.
  - Values = quantities × unit values.
- Consistency strategies among production data, trade data, and Eora:
  - If sum of trade values < aggregate Eora sector value: allow a residual equal to the difference (acts as an ‘other’ category).
  - If sum of trade values > Eora figure: scale down exports or self-trade proportionately, leaving the residual at zero.
  - If export value > production value: set self-trade = 0 (to account for export product classifications possibly incorporating additional value added beyond raw commodities).
- Final dataset coverage (2019):
  - Values (USD) at exporter sector/commodity − importer level including self-trade.
  - 24 aggregated (manufacturing and services) sectors.
  - 82 agricultural commodities.
  - 54 mining commodities.
  - 145 countries.
- Crosswalks: developed by hand to map production data products and HS codes in trade data to the chosen commodity list to link production and trade data through the commodity list.
- Definitions:
  - Self-trade: value of production consumed in the producing country (i.e., not exported).

### Trade elasticities (short vs. long run)
- Role: trade elasticity 휃휃_k measures responsiveness of trade flows to trade barriers; crucial model parameter governing substitution ease and sensitivity of trade flows to international trade costs.
- Baseline long-run elasticities: use sector-level estimates from Fontagné et al. (2022).
  - Example values taken from Fontagné et al. (2022):
    - Mining commodities: -3.41.
    - Agricultural commodities: -2.91.
    - Maximum across sectors reported: -10.56.
  - For services: Fontagné et al. (2022) estimate for one services sector; in counterfactuals all services trade collapsed into one sector.
- Short-run vs. long-run:
  - Boehm et al. (forthcoming) estimate that elasticities can take 7 to 10 years to stabilize; elasticities after one year equal 36 percent of the long-run value.
  - Calibration: ratio between short-run and long-run trade elasticities set in line with Boehm et al. (forthcoming).
  - Note: Andersen and Yotov (2023) estimate short-run elasticity ≈ 10 percent of long-run; authors choose Boehm et al. (higher short-run share) for conservatism.
- Methodological note:
  - Boehm et al. (forthcoming) propose time-differencing the gravity specification reducing long-run elasticity estimates relative to common literature ranges (typical literature: 1.75 to 2.25).
  - Authors opt for Fontagné et al. (2022) long-run estimates to ensure conservative results while applying Boehm et al. short-run/long-run ratio to generate short-run results.
- Data and estimation notes:
  - Fontagné et al. (2022) use bilateral tariffs for the universe of country pairs from 2001-2016, gravity approach with fixed effects.
  - Trade elasticity 휃휃_k appears in the model’s gravity equation for imports:
    - ln X_mn^k = −θ_k ln τ_mn^k − θ_k ln p_m^k + ln D_n^k + ln Φ_n^k.

### Demand elasticities
- Model structure: CES aggregator for commodities and other inputs; other sectors use Cobb-Douglas (elasticity of substitution implicitly one).
- CES elasticity of substitution between inputs: set equal to 0.2 (authors take the conservative side of the mode of demand elasticity estimates surveyed by FS).
- Relationship to price elasticity of demand:
  - If an individual commodity is a small share of total costs (holding other inputs constant), the price elasticity of demand for commodity j ≈ elasticity of substitution η.
  - Approximation: ∂ ln(d_mnkj / p_mnkj) / ∂ ln p_mnkj ≈ η.

### Additional modelling and analytical points (as related to data)
- Model uses exact hat algebra to simulate impacts of changes in trade barriers; static model (no endogenous changes in labor productivity via capital accumulation or technological change).
- First-order approximations derived in the model:
  - Change in local price index due to elimination of trade between two blocs in sector k:
    - d ln P_n^k ≈ (1 / θ_k) ln(1 / (1 − π_n,otherbloc^k))
    - π_n,otherbloc^k = share of initial expenditures of country n on goods from countries in the other bloc.
  - First-order change in output summarized by three terms (all scaled by 1 / θ_k):
    - Direct effect on final goods prices.
    - Indirect effect through input-output linkages.
    - Effect of access to commodities.
  - Equation (6) shows output impacts depend on:
    - 훼_m (1−μ_m) = country’s average expenditure share on intermediate inputs from non-commodity sectors.
    - 훼_m μ_m = country’s average expenditure share on intermediate inputs from commodity sectors.
    - Weighted averages of ln(1−π_n,otherbloc^k) for non-commodity and commodity goods using expenditure shares as weights.
- Model boundaries and assumptions:
  - Does not account for changes in labor productivity due to capital accumulation or technological change—short-run disruptions do not produce long-run labor productivity losses within the model.
  - Initial aggregate trade imbalances are allowed but assumed fixed (do not respond endogenously); robustness checks exogenously vary initial aggregate imbalances.

*Source: wpiea2023073-print-pdf — 2.  Data*

### introduction of barriers that prohibit trade in energy, agricultural goods and ‘high-tech’ sectors respectively.

### wpiea2023073-print-pdf - introduction of barriers that prohibit trade in energy, agricultural goods and ‘high-tech’ sectors respectively.

### Scenario design and modelling assumptions
- High-tech sectors defined using OECD (2011) classification based on sectoral R&D intensities; highlighted high-tech sectors: electronics and machinery, and transport equipment.
- Counterfactual scenarios assume the introduction of infinite trade costs for the specific sectors mentioned, ensuring zero trade occurs between the relevant countries/groups.
- Five illustrative scenarios (summary of assumptions in text table; RoW = rest of the world):
  - Scenario A: Trade barriers on energy sector.
  - Scenario B: Trade barriers on high-tech sectors.
  - Scenario C: Trade barriers on agriculture goods.
  - Strategic decoupling (SD): USA and EU impose barriers to prohibit trade of all goods and services from Russia and vice versa; additional barriers on trade in high-tech sectors between the US-EU and China; no changes in trade relations between US, EU, China, Russia, and RoW.
  - Geo-economic fragmentation (GEF): USA and EU impose barriers on all trade with China; countries from RoW forced to choose exclusive trade with either the US-EU or Russia-China based on historical trade intensities; tariffs on all sectors by the other group.
- Short-run corresponds to one year after barrier increase with a low associated trade elasticity; long-run corresponds to 10 years after the barrier increase with a higher trade elasticity.
- For strategic sectors (commodities and high-tech) the short-run trade elasticity is set to 36 percent of the long-run value.

### Key modelling features emphasised
- Use of a newly constructed granular commodity dataset versus standard aggregated IO (Eora-26) matrix.
- Exact hat algebra algorithm used for estimates (Annex).
- Impacts presented in changes in output, or real GDP; aggregation to group-level and global impacts using real GDP (PPP) weights from 2019.
- Counterfactual autarky experiment: increase all international trade barriers such that there is no international trade in the new equilibrium and set trade imbalances to zero.

### Main quantitative results by scenario (selected figures preserved exactly)
- Granularity importance:
  - Welfare loss more than doubles for LICs when incorporating disaggregated commodity trade versus standard IO.
  - For AEs welfare losses increase by 4 percent; for EMEs welfare losses increase by 25 percent (relative increases from aggregated to disaggregated commodity treatment).
- Scenario A (energy trade eliminated):
  - Short-run output decreases: EMs: 10.1 percent; AEs: 7.8 percent; LICs: 6.3 percent.
  - Long-run impacts range from 1.1 to 2.1 percent across country groups.
- Scenario B (zero high-tech trade):
  - AEs experience the largest output loss (reflects broader representation of high-tech trade for AEs).
- Scenario C (zero agricultural trade):
  - Underscores importance of international trade in agriculture especially for LICs.
- Strategic Decoupling (SD):
  - Short-run output reductions: AEs: 0.7 percent; EMs: 0.8 percent.
  - Long-run output reduction: 0.3 percent for AEs and EMs.
  - LICs benefit slightly over both time periods due to trade diversion.
  - Global output: fall of 0.8 percent in the short run and 0.3 percent in the long run (baseline elasticities).
- Geo-economic Fragmentation (GEF):
  - Aggregate long-run reduction in real GDP: 2.3 percent.
  - Aggregate short-run reduction in real GDP: 4.8 percent.
  - Short-run losses: AEs: 4.2 percent; EMs: 5.2 percent; LICs: 10.8 percent.
  - Long-run losses: AEs: 2.1 percent; EMs: 2.5 percent; LICs: 4.3 percent.
  - LICs most exposed due to: (i) higher commodity intensity (especially agriculture), (ii) smaller size and higher reliance on commodity imports, (iii) specialization in exports of key commodities (metals and energy).
- Regional impacts (baseline, long-run elasticities):
  - SD yields positive impacts in several regions via trade diversion; largest positive for Middle East and Central Asia; largest negative for Emerging Market Europe (losses reduce from 3.8 percent of GDP to 3.0 percent if Russia and Turkey are excluded).
  - GEF produces negative impacts across all regions; Latin America and the Caribbean show the smallest negative impact; Emerging Market Europe shows the largest losses.

### Robustness checks and sensitivity
- Trade elasticity selection:
  - Long-run trade elasticity mean used: 6.00.
  - Literature bounds used for robustness: maximum value 6.6 (Broda and Weinstein, 2006) and minimum value 2.12 (Boehm et al., forthcoming).
  - Geo-fragmentation scenario welfare-cost range: 1.9 percent of GDP (high elasticity case) to 7.0 percent of GDP (low elasticity case), with baseline estimate 2.3 percent of GDP.
- Inter-bloc aggregate imbalances:
  - Re-running baseline with scaled imbalances such that all surpluses and deficits cancel out within each bloc yields results very close to baseline (scaling factors on the order of 15 percent); changes in real GDP are muted for several reasons (nominal wages and prices move together, full employment assumption, no capital flows or technology impacts).
- Alternative country bloc formation (geopolitical ‘closeness’ via Ideal Point Distance, IPD):
  - IPD-based (Geopolitical blocs) assignment increases global aggregate losses to 3.2 percent of global GDP, higher than baseline bloc formation based on trade intensities.
  - Table 3: GDP Losses for Robustness Checks (GEF, long-run elasticities)
    - AE: Baseline -2.14; No inter-bloc imbalances -2.09; Geopolitical blocs -3.40
    - EM: Baseline -2.48; No inter-bloc imbalances -2.36; Geopolitical blocs -3.03
    - LIC: Baseline -4.25; No inter-bloc imbalances -4.12; Geopolitical blocs -4.17
    - Global: Baseline -2.34; No inter-bloc imbalances -2.25; Geopolitical blocs -3.20

### Conclusions and implications highlighted in the chapter
- Not accounting for granular commodity production and trade linkages leads to underestimation of output losses from trade fragmentation; LICs face losses more than twice those estimated with aggregate commodity sectors; AEs and EMs face costs that are 4 percent and 25 percent larger respectively under granular treatment.
- Output losses increase with the depth of fragmentation; LICs are disproportionately affected, especially under severe fragmentation forcing exclusive bloc alignment.
- Short-run adjustment costs (elasticities lower as fraction of long-run values) can be large: example short-run cost for LICs of 10.8 percent when facing adjustment costs that reduce their ability to adapt quickly.
- Baseline loss estimates are conservative relative to ranges in the literature; assumptions on elasticities matter materially.
- Formation of blocs based on geopolitical allegiances (rather than economic trade intensities) raises global losses materially; closing inter-bloc trade imbalances does not materially change results.
- Future research directions noted:
  - Microfound short- and long-run elasticity differences via quadratic adjustment costs in a dynamic system.
  - Introduce endogenous capital formation and technological progress to capture long-run welfare effects from short-run adjustments.
  - Relax exogenous overall trade imbalance assumption.
  - Explore endogenous bloc formation algorithms where countries choose blocs iteratively based on evolving welfare benefits.

*IMF Working Paper excerpt (Fragmentation in Global Trade).*

### References

### References (Content unit: wpiea2023073-print-pdf - References)

### Key bibliographic entries
- Major cited works include:
  - Aiyar, S., Chen, J., Ebeke, C., Garcia-Saltos, R., Gudmundsson, T., Ilyina, A., Kangur, A., Rodriguez, S., Ruta, M., Schulze, T., Trevino, J., Kunaratskul, T. and Soderberg, G. “Geoeconomic Fragmentation and the Future of Multilateralism”. IMF Staff Discussion Note, SDN/2023/01, International Monetary Fund, Washington, DC.
  - Albrizio, S., Bluedorn, J. C., Koch, C., Pescatori, A., & Stuermer, M. (2022). Market Size and Supply Disruptions: Sharing the Pain from a Potential Russian Gas Shut-off to the EU. IMF Working Papers, 2022(143).
  - Amiti, M., Redding, S. J., & Weinstein, D. E. (2019). The impact of the 2018 tariffs on prices and welfare. Journal of Economic Perspectives, 33(4), 187-210.
  - Anderson, J. E., & Yotov, Y. V. (2023). Estimating gravity from the short to the long run: A simple solution to the'International Elasticity Puzzle' (No. w30809). National Bureau of Economic Research.
  - Arkolakis, C., Costinot, A., & Rodríguez-Clare, A. (2012). New trade models, same old gains?. American Economic Review, 102(1), 94-130.
  - (Additional cited authors and works follow in the source list, including Fontagné et al. (2022); Di Bella et al. (2022); Fajgelbaum et al. (2020); Ossa (2014, 2015); Simonovska & Waugh (2014); and many IMF Working Papers and regional outlooks.)

### Annex I — Data: Commodity list
- Commodity list (as provided): AlmondsLettuceAntimonyRare Earths, ApplesLinseedArsenicSalt, ApricotsMaizeAsbestosSelenium, AsparagusMandarinsBarytes & StrontiumSillimanite, AvocadosMangoesBauxiteSilver, BananasMateBerylTalc, BarleyMelonsBismuthTellurium, BeansMilletBoratesTin, BerriesMushroomsBromineTitanium, Brazil nutsMustard seedsCadmiumTungsten, BuckwheatNatural rubberChromiumUranium, Canary seedNutmegCoalVermiculite, CarrotsOatsCobaltWollastonite, CashewsOnionsCopperZinc, CassavaOrangesCrude OilZirconium, CauliflowersPapayasDiamondOther minerals, CherriesPeachesDiatomite, ChestnutsPearsFeldspar, ChickpeasPeasFluorspar, CinnamonPineapplesGallium,indium,rhenium, thallium, ClovesPistachiosGermanium, CocoaPlumsGold, CoconutsPoppy seedsGraphite, CoffeePotatoesGypsum, Cotton seedsRapeseedIodine, CucumbersRiceIron & Steel, DatesRyeLead, EggplantsSorghumLithium, FigsSoy beansMagnesium, FlaxSpinachManganese, GarlicSugarMercury, GingerSunflower seedsMica, GrapefruitSweet potatoesMolybdenum, GrapesTeaNatural Gas, HazelnutsTobaccoNickel, HopsTomatoesNiobium, tantalum, vanadium, JuteVanillaPalladium, LeeksWalnutsPhosphate, LemonsWheatPlatinum, LentilsOther agriculture goodsPotash, Agriculture Commodity, Mineral Product.

### Trade elasticities (adapted table and mapping)
- Source: Adapted from Fontagné et al. (2022), Table 8. Notes preserve original assignment rules and definitions.
- Trade elasticities by sector (Sector – TiVA 2016 Sector(s) - Eora — trade elasticity):
  - Agriculture, hunting, forestry and fishing — Agriculture, Fishing — -2.91
  - Mining and quarrying — Mining and Quarrying — -3.41
  - Food products, beverages and tobacco — Food & Beverages — -4.17
  - Textiles, textile products, leather and footwear — Textiles and Wearing Apparel — -4.71
  - Wood and products of wood and cork / Pulp, paper, paper products, printing and publishing — Wood and Paper — -8.80; -8.21 (assigned as an unweighted average when combined)
  - Coke, refined petroleum products and nuclear fuel / Petroleum, Chemical and Non-Metallic Mineral Products — -3.67
  - Chemicals and chemical products — -10.56
  - Rubber and plastics products — -6.75
  - Other non-metallic mineral products — -4.79
  - Basic metals — Metal Products — -7.39
  - Fabricated metal products — -4.22
  - Machinery and equipment, nec / Computer, electronic and optical equipment / Electrical machinery and apparatus, nec — Electrical and Machinery — -5.01; -5.14; -4.11
  - Motor vehicles, trailers and semi-trailers — Transport Equipment — -8.92
  - Other transport equipment — -8.99
  - Manufacturing nec; recycling — Other Manufacturing, Recyling — -4.06
  - Other community, social and personal services / Construction-Education — -8.35 (assigned for all services sectors)
- Notes preserved:
  - For Eora sectors that comprise multiple TiVA sectors, the unweighted average estimate is assigned (example: Wood and Paper assigned a trade elasticity of -8 .505 — the average of -8.8 and -8.21, shown in source as -8 .505).
  - Trade elasticities are defined as the change in trade in response to a change in trade costs.

### Annex II — Modelling framework details: structure and key equilibrium conditions
- Environment and classification:
  - There are N countries (m = exporter, n = importer), K non-commodity sectors (k), and J commodity sectors (j).
- Production and aggregation:
  - Representative retailer maximizes (A.1) subject to (A.2): composite final good Q_m aggregated from non-commodity inputs Q~_m_k with ∑σ_m_k_k = 1.
  - Non-commodity wholesaler problem: objective (A.3) with constraint (A.4), where x_m m' k are inputs from countries m, θ_k is sector trade elasticity, ∑ω_m m' k_m = 1, and τ_m m' k are trade costs.
  - Variety producer problem for non-commodity sectors: objective (A.5) with production function (A.6). Variables include L_m_k, O_m_k, M_m_k, productivity z_m_k, elasticity η.
  - Commodity retailer aggregates J commodities via (A.7)-(A.8).
  - Commodity-sector wholesaler and producer problems summarized by (A.9)-(A.12) and associated constraints (A.10), (A.11), (A.12).
- Household budget constraint:
  - P_m C_m = w_m L_m + D_m (A.13), with ∑D_n = 0.
- Trade costs and iceberg formulation:
  - Trade subject to iceberg trade costs: shipping one unit of good in sector k from m to n requires τ_m n k ≥ 1 units produced in m, with τ_n n k = 1. Triangular inequality holds: τ_m n k τ_n h k ≥ τ_m h k for all n, m, h.
- Equilibrium definition and aggregate identities:
  - Market clearing and accounting conditions include:
    - Q_m = C_m + ∑M_m_k + ∑M_m_j (A.22)
    - Q_m_k = Q~_m_k (A.23)
    - Q_m_j = Q~_m_j (A.24)
    - L_m = ∑L_m_k + ∑L_m_j (A.25)
    - ∑p_m_k q_m_k + ∑p_m_j q_m_j + D_m = w_m L_m + ∑(P_m M_m_k + P_m O_m_k) + ∑(P_m M_m_j + P_m O_m_j) (A.26)
  - Representative retailer and wholesaler optimality yield price and markup relations such as (A.14)-(A.17).
  - Producer optimality yields labor, input, and marginal price relations (A.18)-(A.21).
- Equilibrium using exact hat algebra (counterfactual relative changes):
  - Hat notation: a variable with a hat represents relative change between counterfactual and observed equilibrium.
  - Key solved relations (examples preserved exactly from source):
    - p̂_m_k = [ζ_m_k P̂_m^O^(1−η) + (1−ζ_m_k)(P̂_m^(α_n_k(1−μ_n_k))...)^(1−η)]^(1/(1−η)) for all non-commodity sectors (formula shown in source).
    - p̂_m_k = ŵ_m^(1−α_n_j) P̂_m^(α_n_j) for all commodity sectors.
    - P̂_m_k = [∑ π_m m' k (τ̂_m m' k p̂_m' k)^−θ_k ]^(−1/θ_k) for non-commodity sectors, and analogously for commodity sectors with θ_j.
    - Aggregate price indexes: P̂_m = ∏(P̂_m_k)^(σ_n_k_k); P̂_m^O = ∏(P̂_m_k)^(σ_j_k_k) (products over sectors with sigma weights).
    - Trade share changes: π̂_m m' k = (τ̂_m n k ⋅ p̂_m_k / P̂_n_k)^−θ_k for non-commodity sectors (and analogous expression for commodity sectors).
    - Total income definition: E_m ≡ w_m L_m + D_m and hat relation Ê_m = D_n E_n + w_n L_n E_n ŵ_m (source text expression preserved).
    - Output and expenditure identities and their hat forms are provided in the source (multi-equation system tying Ŷ, Ê, π̂, and X changes).
- Additional model constructs:
  - Definitions of Ê_m^M and sectoral production Ŷ_m_k as functions of trade flows X_m n k, shares π̂, and income/expenditure ratios are preserved in the source equations.
  - Market clearing in hat algebra: ∑ Y_m_k = ∑ Ŷ_m_k + D_m = E_m ⋅ Ê_m + ∑(P_m M_m_k + P_m O_m_k) Ŷ_m_k + ∑P_m M_m_k Ŷ_m_k (expression preserved in full form in source).

*Fragmentation in Global Trade: Accounting for Commodities — Working Paper No. WP/2023/075*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023073-print-pdf.pdf_