## CHAPTER 3 ONLINE ANNEX — FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES

## Source details

**Canonical URL:** [CHAPTER 3 ONLINE ANNEX — FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES](https://www.imf.org/-/media/files/publications/weo/2023/october/english/ch3onlineannex.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/weo/2023/october/english/ch3onlineannex.pdf.md)
- [Structured JSON version](/-/media/files/publications/weo/2023/october/english/ch3onlineannex.pdf.json)

---

### Baseline scenario and bloc composition
- Baseline divides countries into two hypothetical geopolitical blocs based on the March 2022 UN vote on the war in Ukraine.
- Countries which abstained in the vote are assigned to the China-Russia+ bloc.
- Online Annex Table 3.1.1. lists member economies in each bloc for the baseline scenario.

### Gravity equation exercise: data, specification, and main empirical finding
- Data sources and covariates:
  - Bilateral trade values from BACI and FAO.
  - Geographical distance between most populated cities, contiguity, common language, colonial relationships, colonial history, and current colony status, and WTO membership from CEPII GeoDist.
  - Military alliances from ATOP; similarity measured by Signorino and Ritter’s (1999) s-score s_iijt and military distance MMD_iijt = 1 − s_iijt, normalized so its standard deviation is 1 in every year.
- Core gravity specification (2010–18 data):
  - y_cijt = α_c MMD_iijt + β_c X_iijt + E_cijt + I_cijt + ε_cijt
    - y_cijt is inverse hyperbolic sine of exports of commodity type c from i to j in year t.
    - X_iijt are gravity covariates (distance, contiguity, common language, colonial links, WTO membership).
    - E_cijt and I_cijt are importer-by-year and exporter-by-year fixed effects.
- Robustness specifications:
  - Poisson (one stage): v_cijt = exp(α_c MMD_iijt + β_c X_iijt + E_cijt + I_cijt + ε_cijt).
  - Poisson (two stages) following Hakobyan and others (2023): first estimate undirected propensity δ_cijtt, then in second stage δ_cijt = α_c MMD_iijt + β_c X_iijt + u_cijt.
- Main empirical finding:
  - Across all specifications, military distance is negatively associated with commodity trade flows.
  - The negative effect of geopolitical distance on trade is typically most pronounced for minerals.

### Gravity equation: coefficients on military distance (specifications and exact estimates)
- Specification: Baseline (inverse hyperbolic sine)
  - All Commodities: –0.2306*** (0.0372)
  - Agriculture: –0.2298*** (0.0376)
  - Energy: –0.1532*** (0.0259)
  - Minerals: –0.3789*** (0.0416)
- Specification: Poisson (one stage)
  - All Commodities: –0.1510.1052** (0.1007)
  - Agriculture: –0.4492*** (0.1499)
  - Energy: –0.2449*** (0.0716)
- Specification: Poisson (two stages)
  - All Commodities: –0.1186* (0.0621)
  - Agriculture: –0.0558 (0.0418)
  - Energy: –0.0957 (0.1014)
  - Minerals: –0.1575*** (0.057)
- Note: All specifications include exporter-by-year and importer-by-year fixed effects. Standard errors in parentheses clustered at the importing country level. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

### Multi-country partial equilibrium commodity market model: setup and aggregation
- Model structure follows Alvarez and others (2023); single commodity, multiple countries.
- Country-specific log-linear supply and demand:
  - ln(q_c^s) = η_s ln(p_c) + γ_c^s with η_s > 0
  - ln(q_c^d) = η_d ln(p_c) + γ_c^d with η_d < 0
  - All countries share η_s and η_d; country-specific shifters γ_c^s and γ_c^d.
- Blocs: B ∈ {US‑Europe+, China‑Russia+}.
- Aggregation to bloc level:
  - ln(Q_B^s) = η_s ln(p_B) + γ_B^s, where γ_B^s = ln(Σ_{c∈B} e^{γ_c^s})
  - ln(Q_B^d) = η_d ln(p_B) + γ_B^d, where γ_B^d = ln(Σ_{c∈B} e^{γ_c^d})

### Integrated and fragmented market equilibria (expressions used)
- Integrated market (non-arbitrage): p_US‑Europe+ = p_China‑Russia+ = p_w
  - ln(p_w) = 1/(η_s − η_d) * (Ω_d − Ω_s)
    - Ω_d ≡ ln(e^{γ_US‑Europe+^d} + e^{γ_China‑Russia+^d})
    - Ω_s ≡ ln(e^{γ_US‑Europe+^s} + e^{γ_China‑Russia+^s})
- Fragmented market (no cross-bloc trade):
  - ln(p_B) = (γ_B^d − γ_B^s) / (η_s − η_d)
- Price change from integrated to fragmented equilibrium:
  - ln(p_B) − ln(p_w) = 1/(η_s − η_d) * [ (γ_B^d − γ_B^s) − (Ω_d − Ω_s) ]
  - Under calibration with p_w = 1 and Ω_d − Ω_s = 0:
    - ln(p_B) − ln(p_w) = (γ_B^d − γ_B^s) / (η_s − η_d)

### Consumer, producer, and total surplus changes (formulas)
- Change in consumer surplus for country c:
  - ΔCS_c = − [ (η_d)^{-1} ln(p_c/p_w) + γ_d ]_{p_w}^{p_c}
  - Implemented in closed form: ΔCS_c = − p_w q_{c,w}^d [ (p_c / p_w)^{1+η_d} − 1 ] / (1 + η_d)
- Change in producer surplus for country c:
  - ΔPS_c = p_w q_{c,w}^s [ (p_c / p_w)^{1+η_s} − 1 ] / (1 + η_s)
- Total surplus = Σ_c (ΔCS_c + ΔPS_c)
- Qualitative insight: Surplus changes are larger in countries experiencing larger price changes in commodities they largely consume or produce.

### Calibration details (data years, elasticities, and quantities)
- Calibration matches pre-fragmentation (2019) country and bloc-level trade flows (except crude oil and zirconium use 2018 due to data quality).
- Demand and supply shifters γ_c^d and γ_c^s are calibrated as logs of initial quantities demanded and supplied with p_w = 1.
- Quantities measured as volume in metric tons of commodity content.
- Quantity demanded calibrated as quantity produced minus net exports volume; where net exports > production while production positive, production set equal to net exports for consistency.
- Elasticities η_d and η_s:
  - For energy and agricultural commodities: average of the minimum and maximum short-run price elasticities in Fally and Sayre (2018).
  - For mineral commodities: median of short-run elasticities in Dahl (2020).
  - If commodity-specific estimates unavailable, use average elasticity for commodity type.
  - Details of elasticities per commodity are in Alvarez and others (2023).

### Partial equilibrium results and alternative bloc configurations
- Annex figures summarized in source:
  - Fragmentation-induced price changes in baseline by commodity and bloc.
  - Price effects from individual countries switching blocs (top 15 exporter-induced increases).
  - The 5 commodities that generate largest drops in total bloc-level surplus.
  - Distribution of country-level changes in total surplus across countries and commodities.
- Alternative bloc configurations (Online Annex Table 3.5.1.):
  - Configuration A: similar to April 2023 WEO Chapter 4 — all emerging and developing economies (excluding India, Indonesia and Latin American countries) assigned to China‑Russia+.
  - Configuration B (major trading partner rule): assignment based on 2019 UN Comtrade trade shares (trade more with US+EU → US‑Europe+; trade more with China+Russia → China‑Russia+).
- Bloc configuration A: notable impacts
  - Crude oil: price increases more in US-Europe+ because major oil producers (UAE, Libya, Nigeria, Qatar, Saudi Arabia, Kuwait) are in China-Russia+.
  - Cocoa: price increases in US-Europe+ because Ivory Coast is in China-Russia+.
  - Cobalt: price rises in US-Europe+ because Democratic Republic of Congo is in China-Russia+.
  - Palm oil and manganese: China-Russia+ experiences milder price increases because Malaysia and Thailand assigned to China-Russia+; manganese less vulnerable in China-Russia+ because India assigned to US-Europe+.
  - Total surplus implications:
    - Crude oil and cocoa cause the largest surplus declines in US-Europe+.
    - These commodities imply surplus losses in US-Europe+ between 2.5 and 4.5 percent of GNE.
    - Crude oil also causes surplus declines in China-Russia+ (over 1 percent of GNE), driven by producer surplus declines.
- Bloc configuration B: major reassignments and effects
  - Assigned to US-Europe+: India, Mozambique, South Africa.
  - Assigned to China-Russia+: Australia, Democratic Republic of the Congo, Indonesia, Korea, Malaysia, New Zealand, Philippines, Thailand.
  - Price/vulnerability changes:
    - US-Europe+ faces large price increases in palm oil and cobalt; less vulnerable to fragmentation of graphite, refined platinum, refined palladium.
    - China-Russia+ still faces large price increases of soybean, copper, manganese, zinc, and lead; not large increases for iron ore and lithium (Australia in China-Russia+) or for palm oil.
  - Changes in total economic surplus are larger in China-Russia+ but lower in magnitude relative to baseline.

### Sensitivity to alternative elasticities
- Alternative specification: demand and supply elasticities set to median values among estimates in Fally and Sayre (2018); missing elasticities replaced with average elasticity by broader categories.
- Findings:
  - Ranking of commodity price vulnerability to fragmentation overall in line with baseline.
  - Results on the five largest surplus changes across blocs broadly in line with baseline.

### GMMET model: structure, mineral-sector extensions, and calibration
- Model class and regions:
  - GMMET is a general equilibrium multi-region multi-sector model configured for six regions: (1) United States, (2) European Union, (3) Countries leaning toward the US and the EU, (4) China, (5) Russia, (6) Countries leaning toward China and Russia.
  - Configured to restrict trade between two hypothetical blocs: US‑Europe+ and China‑Russia+.
- Core macro structure:
  - Large-scale structural New-Keynesian dynamic general equilibrium model; each period corresponds to one calendar year.
  - Households: liquidity-constrained households consume all income each period; overlapping-generations households choose consumption, saving, and labor supply.
  - Consumption includes standard goods and services, energy for residential purposes (natural gas and electricity), and transportation services.
  - Transportation: conventional cars (gasoline) and EVs (electricity); vehicle choice depends on relative prices.
  - Firms produce tradable and nontradable goods using energy inputs (fossil and renewable), labor, and capital.
  - Fiscal and monetary rules, nominal and real rigidities, market clearing and CPI indexes included.
- Fossil fuel energy sectors:
  - Energy production from coal, gas, crude oil; each combines capital and labor with a resource fixed in each period but scalable over time.
  - Oil and gas consumed by households; natural gas and coal used for electricity.
  - Oil and coal markets modeled with hedger setting global-clearing prices in integrated markets and bloc-level clearing prices under restrictions.
  - Natural gas modeled as bilateral flows due to pipeline and LNG constraints.
- Minerals sector extensions:
  - Minerals added: Copper, nickel, lithium, cobalt.
  - Mineral composites:
    - Mineral1: copper and nickel.
    - Mineral2: lithium and cobalt.
  - Uses: mineral1 in cables, turbines, panels, conventional cars (CC) and renewables; mineral2 in EV batteries.
  - Production structure: CCs and EVs combine mineral1 with investment goods; EVs additionally require mineral2. Renewable structures use mineral1.
  - Focus on mining stage; minerals tradable in integrated and segmented markets analogous to oil/coal.
- Data and calibration:
  - Technologies: CES with constant returns to scale.
  - Calibrated to reproduce empirical supply elasticities of fossil fuels and four critical minerals (Fally and Sayre, 2018; Dahl, 2020).
  - Trade and production intensities calibrated using BGS, US Geological Survey, Bilateral Commodity Trade Database, IEA, Eurostat.
  - Extraction and use primarily based on 2018 and 2019 data; copper and cobalt use 2018 data; lithium 2019; nickel average 2015–2019.
- Mineral use, supply and trade (percent of region GDP — exact table values)
  - Regions correspond to: (1) United States, (2) European Union, (3) Countries leaning toward USA-Europe+, (4) China-Russia, (5) Countries leaning toward China-Russia+, (6) (unnamed)
  - GDP (percent of world): 24.60 18.10 27.60 16.50 1.80 11.40
  - Mineral1 (copper and nickel): 0.03 0.07 0.28 0.41 0.61 0.10
  - Mineral1 Production: 0.04 0.04 0.48 0.08 0.53 0.16
  - Mineral1 Net imports: –0.01 0.03 –0.19 0.33 0.07 –0.06
  - Mineral2 (lithium and cobalt): 0.00 0.01 0.01 0.03 0.02 0.01
  - Mineral2 Production: 0.00 0.00 0.03 0.00 0.01 0.00
  - Mineral2 Net imports: 0.00 0.01 –0.02 0.03 0.01 0.00
  - Note: Accounting errors due to rounding. Minerals data are at the mined stage.
- Minerals’ elasticities of substitution (benchmark and higher)
  - Minerals and other factors in manufacturing: 0.2 / 0.4
  - Minerals in the production of electric transport: 0.2 / 0.4
  - Mineral1 in the production of conventional transport: 0.2 / 0.4
  - Mineral1 and production of renewables: 0.1 / 0.4
  - Benchmark assumes low substitutability for mineral1 in renewables (0.1); robustness tested with 0.4.

### Fragmentation scenarios: implementation, channels, and mechanics
- Implementation:
  - Model starts from steady-state with fully integrated markets.
  - Fragmentation introduced by eliminating trade in each key commodity between the two blocs.
  - For gas, fragmentation restricts bilateral trade between regions in opposite blocs (trade modeled as bilateral flows).
- Main propagation channels:
  - Trade diversion: trade reallocated within blocs; hedger reallocates trade so supply and demand clear at bloc-level prices.
  - Price channel: ex-ante net exporting bloc experiences price declines; ex-ante net importing bloc experiences price increases.
  - Rigidities: pipelines, processing/refining capacity limit near-term trade diversion, slowing volume adjustment and amplifying price and aggregate output effects.

### Comparison: crude oil versus natural gas fragmentation (key quantitative outcomes)
- Crude oil fragmentation (impact year 1):
  - Oil prices: increase by about 18 percent in the US-Europe+ bloc and decline by about 28 percent in the China-Russia+ bloc.
  - Traded oil volumes adjust substantially in first year with contained GDP and inflation impacts.
  - Exports: oil exports increase within US-Europe+ and decrease in China-Russia+.
  - Russia faces largest GDP losses as oil exports represent about 50 percent of its total exports.
  - European Union faces larger losses within US-Europe+ as a net importer in a bloc with price increases.
- Natural gas fragmentation:
  - Impact on GDP and inflation more marked and negative in both blocs due to rigidities (pipelines, LNG infrastructure).
  - China imports more than 60 percent of its gas from countries belonging to the opposite bloc that cannot be quickly replaced.
  - Result: larger impacts on trade flows and inflation and marked negative GDP impacts in near-term for the European Union and China.
  - Russia faces more pronounced losses relative to oil due to rigidities.

### Fragmentation of critical minerals markets and implications for the clean energy transition
- Mechanisms and constraints:
  - Elevated concentration of mined minerals supply in US-Europe+ leads to steep price increases and inflation in China-Russia+ when fragmentation occurs.
  - Heavy use of mineral1 in manufacturing and construction in countries like China generates a large fall in GDP in China-Russia+.
  - Oversupply in US-Europe+ cannot be quickly used domestically because processing and refining capacity requires 5 to 10 years to build and scale up.
  - Model imposes a shock to productive use of minerals in US-Europe+ proportional to China-Russia+ initial import share; shock diminishes over 10 years.
- Simulation of green-transition demand:
  - Assumes demand increases for copper, nickel, cobalt, lithium as projected by IEA (2023) Net Zero Emission scenario (NZE baseline).
  - Increase in investment in renewable energy and EVs simulated via “green” subsidies in all regions through 2030.
  - Baseline: minerals traded freely and subsidies present.
  - Fragmentation scenario: bans minerals trade across blocs while leaving subsidies unchanged.
- Effects under fragmentation:
  - Initial years: mineral prices increase in both blocs because China-Russia+ cannot access minerals mined in US-Europe+; US-Europe+ cannot immediately use excess due to limited refining capacity.
  - China-Russia+ experiences steep price increases, inflation, and large GDP declines from reduced availability and higher costs of mineral inputs.
  - US-Europe+ experiences GDP decline because a share of minerals that cannot be exported cannot be used domestically until refining capacity is built (5 to 10 years).
  - Time path: productive-use shock proportional to initial import share diminishes over 10 years to reflect 5 to 10 years to set up refinery plants.

### Key findings on mineral prices, investment, and transition outcomes (quantitative highlights)
- By 2030, prices are over 20 percent lower in the US-Europe+ bloc and over 200 percent higher in the China-Russia+ bloc, relative to the NZE baseline.
- On the whole NZE transition path 2023-2030, prices of minerals are on average 300 percent higher on average relative to the NZE baseline in the China-Russia+ bloc, accounting for the spike in prices in the initial years.
- Prices fall over time consistent with oversupply of minerals in the US-Europe+ bloc due to greater refining capacity there.
- Fragmentation leads to a decline in investment in renewable energy and EVs at the global level, with much bigger losses, relative to the baselines, in the China-Russia+ bloc.
- If production exhibited increasing returns to scale, the US-Europe+ bloc could scale up investment faster than in the baseline; in the benchmark (constant returns) both blocs see declines in renewables and EVs magnitude as assumed.

*Source: IMF staff compilation; CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES, International Monetary Fund | October 2023*

### CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES

### CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES

### Baseline scenario and bloc composition
- Baseline divides countries into two hypothetical geopolitical blocs based on the March 2022 UN vote on the war in Ukraine (see Online Annex Table 3.1.1.).
- Countries which abstained in the vote are assigned to the China-Russia+ bloc.
- Online Annex Table 3.1.1. lists member economies in each bloc for the baseline scenario.

### Gravity equation exercise: data, specification, and main finding
- Data sources and covariates:
  - Bilateral trade values from BACI and FAO.
  - Geographical distance between most populated cities, contiguity, common language, colonial relationships, colonial history, current colony status, and WTO membership from CEPII GeoDist.
  - Military alliances from ATOP; similarity measured by Signorino and Ritter’s (1999) s-score s_iijt and military distance MMD_iijt = 1 − s_iijt, normalized so its standard deviation is 1 in every year.
- Core gravity specification (2010–18 data):
  - y_cijt = α_c MMD_iijt + β_c X_iijt + E_cijt + I_cijt + ε_cijt
    - y_cijt is inverse hyperbolic sine of exports of commodity type c from i to j in year t.
    - X_iijt are gravity covariates (distance, contiguity, common language, colonial links, WTO membership).
    - E_cijt and I_cijt are importer-by-year and exporter-by-year fixed effects.
- Robustness specifications:
  - Poisson (one stage): v_cijt = exp(α_c MMD_iijt + β_c X_iijt + E_cijt + I_cijt + ε_cijt).
  - Poisson (two stages) following Hakobyan and others (2023): first estimate undirected propensity δ_cijtt, then in second stage δ_cijt = α_c MMD_iijt + β_c X_iijt + u_cijt.
- Main empirical finding:
  - Across all specifications, military distance is negatively associated with commodity trade flows.
  - The negative effect of geopolitical distance on trade is typically most pronounced for minerals.

### Gravity equation: coefficients on military distance (Online Annex Table 3.3.1.)
- Specification: Baseline (inverse hyperbolic sine)
  - All Commodities: –0.2306*** (0.0372)
  - Agriculture: –0.2298*** (0.0376)
  - Energy: –0.1532*** (0.0259)
  - Minerals: –0.3789*** (0.0416)
- Specification: Poisson (one stage)
  - All Commodities: –0.1510.1052** (0.1007)
  - Agriculture: –0.4492*** (0.1499)
  - Energy: –0.2449*** (0.0716)
- Specification: Poisson (two stages)
  - All Commodities: –0.1186* (0.0621)
  - Agriculture: –0.0558 (0.0418)
  - Energy: –0.0957 (0.1014)
  - Minerals: –0.1575*** (0.057)
- Note: All specifications include exporter-by-year and importer-by-year fixed effects. Standard errors in parentheses clustered at the importing country level. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.

### Multi-country partial equilibrium commodity market model: setup
- Model structure (single commodity, multiple countries) follows Alvarez and others (2023).
- Country-specific supply and demand (log-linear):
  - ln(q_c^s) = η_s ln(p_c) + γ_c^s
  - ln(q_c^d) = η_d ln(p_c) + γ_c^d
    - c denotes country.
    - q_c^s and q_c^d are quantities supplied and demanded.
    - p_c is the commodity price faced by country c.
    - η_s > 0 (supply price elasticity), η_d < 0 (demand price elasticity).
    - All countries share η_s and η_d but have unique shifters γ_c^s and γ_c^d.
- Blocs: countries are in one of two blocs B ∈ {US‑Europe+, China‑Russia+}.
- Aggregation to bloc level:
  - ln(Q_B^s) = η_s ln(p_B) + γ_B^s
  - ln(Q_B^d) = η_d ln(p_B) + γ_B^d
  - Q_B^s = Σ_{c∈B} q_c^s ; Q_B^d = Σ_{c∈B} q_c^d
  - γ_B^s = ln(Σ_{c∈B} e^{γ_c^s}) ; γ_B^d = ln(Σ_{c∈B} e^{γ_c^d})

### Integrated market equilibrium
- Market clearing and non-arbitrage conditions:
  - Q_US‑Europe+^s + Q_China‑Russia+^s = Q_US‑Europe+^d + Q_China‑Russia+^d
  - p_US‑Europe+ = p_China‑Russia+ = p_w (world price)
- World price expression:
  - ln(p_w) = 1/(η_s − η_d) * (Ω_d − Ω_s)
    - Ω_d ≡ ln(Σ e^{γ_c^d} across blocs) = ln(e^{γ_US‑Europe+^d} + e^{γ_China‑Russia+^d})
    - Ω_s ≡ ln(Σ e^{γ_c^s} across blocs) = ln(e^{γ_US‑Europe+^s} + e^{γ_China‑Russia+^s})

### Fragmented market equilibrium (no trade between blocs)
- Internal trading costs within a bloc are zero; no cross-bloc trade.
- Bloc-level equilibrium price:
  - ln(p_B) = (γ_B^d − γ_B^s) / (η_s − η_d)
- Bloc- and country-level quantities and net exports derived by substituting p_B into supply and demand curves.

### Fragmentation impact (price change)
- Price change from integrated to fragmented equilibrium:
  - ln(p_B) − ln(p_w) = 1/(η_s − η_d) * [ (γ_B^d − γ_B^s) − (Ω_d − Ω_s) ]
- Under calibration where p_w is standardized to 1 and initial Ω_d = Ω_s (so Ω_d − Ω_s = 0):
  - ln(p_B) − ln(p_w) = (γ_B^d − γ_B^s) / (η_s − η_d)

### Consumer, producer, and total surplus changes
- Change in consumer surplus for country c:
  - ΔCS_c = − [ (η_d)^{-1} ln(p_c/p_w) + γ_d ]_{p_w}^{p_c}
  - As implemented in closed form:
    - ΔCS_c = − p_w q_{c,w}^d [ (p_c / p_w)^{1+η_d} − 1 ] / (1 + η_d)  [expression in source text]
- Change in producer surplus for country c:
  - ΔPS_c = p_w q_{c,w}^s [ (p_c / p_w)^{1+η_s} − 1 ] / (1 + η_s)  [expression in source text]
- Where p_c = p_B in fragmented equilibrium, and p_w q_{c,w}^d and p_w q_{c,w}^s are quantities in dollars demanded and supplied in the integrated equilibrium.
- Total surplus = ΔCS_c + ΔPS_c.
- Key qualitative insight: Surplus changes are larger in countries experiencing larger price changes in commodities they largely consume or produce.

### Calibration details
- Calibration matches pre-fragmentation (2019) country and bloc-level trade flows (except crude oil and zirconium use 2018 due to data quality).
- Demand and supply shifters γ_c^d and γ_c^s are calibrated as logs of initial quantities demanded and supplied with p_w = 1.
- Quantities measured as volume in metric tons of commodity content.
- Quantity demanded calibrated as quantity produced minus net exports volume; where net exports > production while production positive, production set equal to net exports for consistency.
- Elasticities η_d and η_s:
  - For energy and agricultural commodities: use the average of the minimum and maximum short-run price elasticities in Fally and Sayre (2018).
  - For mineral commodities: use the median of short-run elasticities in Dahl (2020).
  - If commodity-specific estimates unavailable, use average elasticity for commodity type.
  - Details of elasticities per commodity are in Alvarez and others (2023).

### Additional partial equilibrium results (summary of contents)
- Annex Figure 3.5.1.: fragmentation-induced price changes in baseline scenario for each commodity in the two blocs.
- Annex Figure 3.5.2.: price effects from individual countries switching blocs; for each bloc, shows the 15 largest price increases induced by an exporter switching trade allegiances.
- Annex Figure 3.5.3.: the 5 commodities that generate the largest drops in total bloc-level surplus in each bloc.
- Annex Figure 3.5.4.: distribution of country-level changes in total surplus from fragmentation across countries and commodities.

### Alternative bloc configurations (Online Annex Table 3.5.1.)
- Two alternatives examined:
  - Bloc configuration A: similar to April 2023 WEO Chapter 4 — all emerging and developing economies (excluding India, Indonesia and Latin American countries) assigned to China‑Russia+; table lists member economies for each bloc under configuration A.
  - Bloc configuration B (main trading partner rule): a country assigned to US‑Europe+ if it trades more with the US and the EU combined than with China and Russia combined; assigned to China‑Russia+ if it trades more with China and Russia combined than with the US and EU combined. Trade shares calculated using 2019 UN Comtrade data.
- Note: the single-commodity partial equilibrium exercise cannot accommodate neutral country assignments used in April 2023 Chapter 4.

*Source: IMF staff compilation; CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES, International Monetary Fund | October 2023*

### CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES

### CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES

### Bloc configuration A: price and surplus impacts
- Price effects (relative to baseline):
  - Crude oil: price increases more in the US-Europe+ bloc because major oil producers are now in the China-Russia+ bloc (UAE, Libya, Nigeria, Qatar, Saudi Arabia, Kuwait).
  - Cocoa: price increases in the US-Europe+ bloc because Ivory Coast, the largest world producer of cocoa, is part of the China-Russia+ bloc.
  - Cobalt: price rises in the US-Europe+ bloc because the Democratic Republic of Congo, the world largest producer of cobalt, has been moved to the China-Russia bloc.
  - Palm oil and manganese: China-Russia+ bloc experiences milder price increases because important palm oil producers such as Malaysia and Thailand are assigned to the China-Russia+ bloc; manganese becomes less vulnerable in China-Russia+ because India is now assigned to the US-Europe+ bloc.
- Total surplus implications:
  - Crude oil and cocoa cause the largest surplus declines in the US-Europe+ bloc.
  - These commodities imply surplus losses in the US-Europe+ bloc between 2.5 and 4.5 percent of GNE.
  - Crude oil also causes relevant surplus declines in the China-Russia bloc (over 1 percent of GNE), driven by producer surplus declines as exporting countries in this bloc experience large reductions in prices.

### Bloc configuration B (major trade partner): reassignments and effects
- Major regional reassignments relative to baseline:
  - Assigned to US-Europe+ bloc: India, Mozambique, South Africa.
  - Assigned to China-Russia+ bloc: Australia, Democratic Republic of the Congo, Indonesia, Korea, Malaysia, New Zealand, Philippines, Thailand.
- Price and vulnerability changes:
  - US-Europe+ bloc: large price increases in palm oil (Indonesia and Malaysia account for 80 percent of global production) and cobalt (DRC shift); less vulnerable to fragmentation of graphite, refined platinum, refined palladium (Mozambique and South Africa now in US-Europe+).
  - China-Russia+ bloc: still experiences large price increases of soybean, copper, manganese, zinc, and lead, but not of iron ore and lithium (Australia assigned to China-Russia+), or of palm oil.
- Changes in total economic surplus:
  - Changes are larger in the China-Russia+ bloc (Figure 3.5.8 in source), but lower in magnitude relative to the baseline.

### Sensitivity to alternative elasticities
- Alternative elasticity specification:
  - Demand and supply elasticities set as the median values among estimates in Fally and Sayre (2018); missing elasticities replaced with average elasticity by broader categories (see Alvarez and others, 2023 for alternatives).
- Findings:
  - Ranking of commodity price vulnerability to fragmentation is overall in line with the baseline (compare Figure 3.5.9 to baseline Figure 3.5.1).
  - Results on the five largest surplus changes across blocs are broadly in line with the baseline (compare Figure 3.5.10 to baseline Figure 3.5.2).

### GMMET model: structure and extensions
- Model class and regions:
  - GMMET is a general equilibrium multi-region multi-sector model configured for six regions: The US; the EU; countries leaning toward the US and the EU; China; Russia; countries leaning toward China and Russia.
  - Configured to restrict trade between two hypothetical blocs: the US-Europe+ bloc and the China-Russia+ bloc.
- Core macroeconomic structure:
  - Belongs to large-scale structural New-Keynesian dynamic general equilibrium models; core described in Carton and others (2023) and Kumhof and others (2010).
  - Time period: each period corresponds to a calendar year.
  - Households: liquidity-constrained households consume all income each period; overlapping-generations households choose consumption, saving, and labor supply.
  - Consumption: standard goods and services, energy for residential purposes (natural gas and electricity), and transportation services.
  - Transportation: conventional cars burning gasoline (from oil) and electric vehicles (EVs) running on electricity; vehicle choice depends on relative prices.
  - Firms: non-energy sectors produce tradable and nontradable goods using energy inputs (fossil and renewable), labor, and capital.
  - Fiscal/monetary: governments and central banks follow specific budgetary and monetary rules; nominal and real rigidities produce notable near-to-medium-term policy effects.
  - Market clearing: all markets clear each period and prices are reflected in CPI indexes in each region.
- Fossil fuel energy sectors:
  - Energy production from three fossil fuel mining sectors: coal, gas, crude oil; each combines capital and labor with a resource fixed in each period but scalable over time.
  - Oil and gas also consumed by households; natural gas and coal used for electricity generation.
  - Oil: domestic and international markets; model allows international oil trade in perfectly integrated markets and in segmented markets.
  - Hedger: aggregates oil exports from producing regions and sells to importing regions; in absence of restrictions hedger sets global-clearing prices; under trade restrictions hedger sets bloc-level clearing prices.
  - Coal markets operate similarly; natural gas markets modeled as bilateral flows due to pipeline and LNG terminal constraints.

### Minerals sector extensions and calibration
- Minerals added: Copper, nickel, lithium, cobalt.
  - Uses:
    - Copper: cables, conductors, essential for turbines and panels in solar and wind generation.
    - Nickel: bearings, shafts, gears, hydraulic components of wind turbines.
    - Lithium and cobalt: essential for lithium-ion batteries in EVs.
- Mineral composites:
  - Mineral1: copper and nickel.
  - Mineral2: lithium and cobalt.
  - Mineral composites used in: manufacturing and construction (tradables bundle), conventional cars (CC), EV batteries (mineral2), and renewable structures.
  - Production structure:
    - CCs and EVs: mineral1 combined with an investment good; EVs additionally require mineral2 as proxy for battery component.
    - Renewable structures: produced with investment goods and mineral1.
    - Tradables: combine mineral1 and mineral2 with capital/labor and energy bundle.
  - Focus on mining stage; mining combines capital and labor with a resource fixed in each period.
  - Minerals tradable in integrated international markets and segmented markets, following oil/coal structure.
- Data and calibration:
  - Technologies: CES with constant returns to scale.
  - Calibrated to reproduce empirical supply elasticities of fossil fuels and four critical minerals (Fally and Sayre, 2018; Dahl, 2020).
  - Trade and production intensities calibrated using BGS, US Geological Survey, Bilateral Commodity Trade Database, IEA, Eurostat.
  - Extraction and use of mineral1 and mineral2 primarily based on 2018 and 2019 data.
  - Note on data: copper and cobalt use 2018 data; lithium data from 2019; nickel uses average exports and production 2015–2019 due to volatility.
- Mineral use, supply and trade (Online Annex Table 3.6.1 — Percent of region GDP, unless noted otherwise):
  - Columns correspond to regions: (1) United States, (2) European Union, (3) Countries learning toward USA-Europe+, (4) China-Russia, (5) Countries leaning toward China-Russia+, (6) (unnamed in table—values listed).
  - GDP (percent of world): 24.60 18.10 27.60 16.50 1.80 11.40
  - Mineral1 (copper and nickel): 0.03 0.07 0.28 0.41 0.61 0.10
  - Mineral1 Production: 0.04 0.04 0.48 0.08 0.53 0.16
  - Mineral1 Net imports: –0.01 0.03 –0.19 0.33 0.07 –0.06
  - Mineral2 (lithium and cobalt): 0.00 0.01 0.01 0.03 0.02 0.01
  - Mineral2 Production: 0.00 0.00 0.03 0.00 0.01 0.00
  - Mineral2 Net imports: 0.00 0.01 –0.02 0.03 0.01 0.00
  - Sources: British geological survey; Gaulier and Zignano (2010); Global macroeconomic model for the energy transition; and IMF staff calculations. Note: Accounting errors due to rounding. Minerals data are at the mined stage.
- Minerals’ elasticities of substitution (Online Annex Table 3.6.2):
  - Benchmark / Higher
  - Elasticity between:
    - Minerals and other factors in manufacturing: 0.2 / 0.4
    - Minerals in the production of electric transport: 0.2 / 0.4
    - Mineral1 in the production of conventional transport: 0.2 / 0.4
    - Mineral1 and production of renewables: 0.1 / 0.4
  - Benchmark assumption: use of mineral1 in production of renewables elasticity = 0.1; robustness tested with higher elasticity = 0.4 to capture greater substitutability.

### Fragmentation scenarios: main channels and mechanics
- Implementation:
  - Model starts from steady-state with fully integrated markets.
  - Fragmentation introduced by eliminating trade in each key commodity between the two blocs.
  - For gas, fragmentation restricts bilateral trade between regions in opposite blocs (trade modeled as bilateral flows).
- Main propagation channels:
  - Trade diversion: trade reallocated within blocs; hedger reallocates trade so supply and demand clear at bloc-level prices.
  - Price channel: ex-ante net exporting bloc (higher initial supply relative to demand) experiences price declines for that commodity; ex-ante net importing bloc experiences price increases.
  - Rigidities: near-term trade diversion may be limited by rigidities (e.g., pipelines, processing/refining capacity), slowing adjustment of trading volumes and amplifying effects on prices and aggregate output.

### Comparison: crude oil versus natural gas fragmentation
- Crude oil fragmentation (impact year 1):
  - Oil prices: increase by about 18 percent in the US-Europe+ bloc and decline by about 28 percent in the China-Russia+ bloc (see Figure 3.6.3 in source).
  - Demand/supply: initial oil demand larger than supply in US-Europe+; opposite in China-Russia+.
  - Dynamics: prices slightly decline as production and trade adjust; adjustment is relatively fast as no material frictions to oil trade.
  - Traded oil volumes: adjust substantially in first year with contained GDP and inflation impacts.
  - Exports: oil exports increase within the US-Europe+ bloc and decrease in the China-Russia+ bloc.
  - Regional impacts:
    - Ex-ante exporters in an ex-ante exporting bloc (e.g., Russia, some countries leaning toward China-Russia+ such as Iraq, Iran) lose because oil prices decline.
    - Russia faces largest GDP losses as oil exports represent about 50 percent of its total exports.
    - China and other mostly oil importers in same bloc benefit from cheaper oil.
    - In US-Europe+ bloc, the European Union faces larger losses as it is a net importer in an ex-ante net importing bloc where prices increase.
- Natural gas fragmentation:
  - Impact on GDP and inflation is more marked and negative in both blocs due to rigidities (pipelines, LNG infrastructure) that limit trade diversion.
  - Examples of rigidities:
    - Gas supplies from Russia to Europe cannot be quickly redirected.
    - China imports more than 60 percent of its gas from countries belonging to the opposite bloc (e.g., Australia) that cannot be quickly replaced with Russian supplies.
  - Result: larger impacts on trade flows and inflation and a marked negative impact on GDP in the near-term in both the European Union and China.
  - Russia: faces more pronounced losses relative to oil due to greater role of rigidities.

### Fragmentation of critical minerals markets and implications for the clean energy transition
- Key mechanisms and constraints:
  - Elevated concentration of mined minerals supply in the US-Europe+ bloc leads to steep price increases and inflation in the China-Russia+ bloc when fragmentation occurs.
  - Heavy use of minerals, especially mineral1, in manufacturing and construction in countries like China generates a large fall in GDP in the China-Russia+ bloc.
  - Oversupply in US-Europe+ cannot be quickly used domestically because processing and refining capacity requires 5 to 10 years to build and scale up.
  - Model captures rigidity by imposing a shock to productive use of minerals in US-Europe+ proportional to China-Russia+ imports of minerals from US-Europe+ in initial equilibrium; this shock gradually diminishes bringing productive use to full supply by year 10.
- Simulation of green-transition demand:
  - Assumes demand for copper, nickel, cobalt, lithium increases as projected by IEA (2023) Net Zero Emission scenario (NZE baseline).
  - Increase in investment in renewable energy and EVs consistent with higher demand up to 2030 simulated via “green” subsidies in all regions.
  - Baseline: minerals traded freely and subsidies present.
  - Fragmentation scenario: bans minerals trade across blocs while leaving subsidies unchanged.
- Effects under fragmentation:
  - Initial years: mineral prices increase in both blocs because China-Russia+ cannot access minerals mined in US-Europe+; US-Europe+ cannot immediately use excess minerals due to limited refining capacity.
  - China-Russia+ experiences steep price increases, inflation, and large GDP declines driven by reduced availability and higher costs of mineral inputs.
  - US-Europe+ also experiences a decline in GDP because a share of minerals that cannot be exported to China-Russia+ after fragmentation cannot be used domestically until refining capacity is built.
  - Time path: shock to productive use mimics difficulty of using minerals without prior refining capacity; shock proportional to initial import share and diminishes over 10 years to reflect 5 to 10 years required to set up refinery plants.

*International Monetary Fund | October 2023 — CHAPTER 3 Online Annex (ch3onlineannex)*

### CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES

### CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES

### Key findings on mineral prices and bloc divergences
- By 2030, prices are over 20 percent lower in the US-Europe+ bloc and over 200 percent higher in the China-Russia+ bloc, relative to the NZE baseline.
- On the whole NZE transition path 2023-2030, prices of minerals are on average 300 percent higher on average relative to the NZE baseline in the China-Russia+ bloc, accounting for the spike in prices in the initial years.
- Prices fall consistent with an oversupply of minerals in the US-Europe+ bloc due to greater refining capacity there.

### Investment and technology impacts
- Fragmentation leads to a decline in investment in renewable energy and EVs at the global level, with much bigger losses, relative to the baselines, in the China-Russia+ bloc.
- A key assumption is the constant-return-to-scale technology in the production of renewable energy and EVs.
- If production exhibited increasing returns to scale, the US-Europe+ bloc could scale up investment faster than in the baseline.
- Both effects deliver a decline in renewables and EVs in the US-Europe+ bloc of the magnitude assumed in the benchmark model.

### Analytical implications and scenarios
- Differential refining capacity and bloc-specific supply/demand dynamics can produce large, divergent price outcomes across blocs during the NZE transition path 2023-2030.
- Initial-year price spikes in the China-Russia+ bloc drive the high average price outcomes reported for that bloc over 2023-2030.

### References (selected from source)
- Alvarez, Jorge, Alexandre Balduino Sollaci, Mehdi Benatiya Andaloussi, Chiara Maggi, Martin Stuermer, and Petia Topalova. 2023. “Geoeconomic Fragmentation and Commodity Markets.” IMF Working Paper, No. 23/201, Washington, DC.
- Blanchard, Olivier. 1985. “Debts, Deficits and Finite Horizons.” Journal of Political Economy 93(21): 223-47.
- British Geological Survey. 2022. “World Mineral Statistics.” Keyworth, Nottingham.
- Carton, B., C. Evans, D. Muir, and S. Voigts 2023 “Getting to Know GMMET The Global Macroeconomic Model for the Energy Transition,” IMF Working Paper, Washington, DC, forthcoming.
- Dahl, Carol. 2020. “Mineral Elasticity of Demand and Supply Database.” Colorado School of Mines Working Paper, No. 2020-2, Golden, Colorado.
- Deutsche Rohstoffagentur (DERA). 2023. Angebotskonzentration bei mineralischen Rohstoffen und Zwischenprodukten – potenzielle Preis- und Lieferrisiken. Report, Deutsche Rohstoffagentur, Spandau, Germany.
- Food and Agriculture Organization of the United Nations. 2023. FAOSTAT Statistical Database. Rome, Italy.
- Fally, Thibault and James Sayre. 2018. “Commodity Trade Matters.” NBER Working Paper, No. 24965.
- Gaulier, Guillaume, and Soledad Zignago. 2010. “BACI: International Trade Database at the Product-Level. The 1994-2007 Version.” CEPII Working Paper, No. 2010-23.
- Hakobyan, Shushanik, Sergii Meleshchuk, and Robert Zymek. 2023. “Divided We Fall: Differential Exposure to Geopolitical Fragmentation and Trade.” IMF Working Paper, Washington, DC, forthcoming.
- International Energy Agency. 2022. “World Energy Statistics.” Paris, France.
- International Energy Agency. 2022. “World Energy Balances.” Paris, France.
- Kumhof, M., D. Laxton, D. Muir, and S. Mursula. 2010. “The Global Integrated Monetary and Fiscal Model (GIMF) – Theoretical Structure.” IMF Working Paper, No 10/34. Washington, DC.
- Jacks, David, and John Tang. 2018. “Trade and Immigration, 1870-2010.” NBER Working Paper No. 25010, Cambridge, MA.
- Jordà, Òscar, Moritz Schularick, and Alan M. Taylor. 2017. “Macrofinancial History and the New Business Cycle Facts.” in NBER Macroeconomics Annual 2016 (31), eds Martin Eichenbaum and Jonathan A. Parker. Chicago: University of Chicago Press.
- Leeds, Brett Ashley, Jeffrey M. Ritter, Sara McLaughlin Mitchell, and Andrew G. Long. 2002. “Alliance Treaty Obligations and Provisions, 1815-1944.” International Interactions 28: 237-260.
- Mayer, Thierry, and Soledad Zignago. 2011. “Notes on CEPII’s distances measures: The GeoDist database.” CEPII Working Paper 2011-25.
- Signorino, Curtis S., and Jeffrey M. Ritter. 1999. “Tau-b or not tau-b: Measuring the similarity of foreign policy positions.” International Studies Quarterly, 43.1: 115-144.
- U.S. Geological Survey. 2021. “Minerals Yearbook.” Reston, Virginia.
- U.S. Geological Survey. 2022. “Mineral Commodity Summaries.” Reston, Virginia.

*Source: CHAPTER 3  FRAGMENTATION AND COMMODITY MARKETS: RISKS AND VULNERABILITIES (excerpt).*

---


_Source: https://www.imf.org/-/media/files/publications/weo/2023/october/english/ch3onlineannex.pdf_