1 Introduction

With Russia’s invasion of Ukraine, individual commodities such as crude oil, natural gas, and wheat were used to exert pressure in a major conflict for the first time since the end of the Cold War. New restrictions on commodities trade spiked in 2022 (Figure 1) contributing to inflationary pressures and affecting output in many countries.

View Full Size

New Trade Restrictions (Index, 2016–2019=100)

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: Global Trade Alert (GTA) and authors’ calculations. This chart shows the number of new trade interventions imposed on commodity sectors in each year. Data is sourced from the GTA database and adjusted for reporting lags (i.e., only interventions reported before Dec 31^st each year are included in the database). Only interventions that the GTA deems “certain to discriminate against foreign commercial interests” are included.

• Download Figure
• Download figure as PowerPoint slide

While most commodity prices normalized in 2023, geopolitical tensions signal that more severe geoeconomic fragmentation¹ of commodity trade is still a major risk. But which commodities are most vulnerable in the event of further disruptions in international trade?

And in which commodity markets would further fragmentation lead to sizable economic losses? In this paper, we develop a comprehensive dataset and framework to assess the relative sensitivity of a broad set of commodity markets to fragmentation and gauge potential economic impacts.

Our analysis proceeds in two steps. First, we construct a new dataset linking production and bilateral trade flows for 48 of the most important energy, mineral and agricultural commodities. Second, we develop a single-commodity partial equilibrium trade model to examine the potential impacts of fragmentation on trade flows, prices, and production of each commodity. Applying this simple framework to individual commodities allows us to assess the relative price sensitivity and macro-importance of a large number of commodities with minimal data requirements.²

After calibrating the model using data from each commodity market in our dataset, we simulate a hypothetical fragmentation scenario in which the global market for each commodity splits into two theoretical blocs. As an illustrative example, we use the 2022 United Nations vote on Russia’s war in Ukraine to classify countries in blocs.³ For each commodity, we assess the implied price change in the two blocs, assuming in a highly stylized way that there is no trade in a particular commodity between blocs, while intra-bloc trade costs are unaffected. The change in commodity prices relative to the price in the fully integrated world is used as a measure of commodity-specific vulnerability in the event of fragmentation.

We then investigate the economic consequences of fragmenting individual commodity markets. We compute the change in total economic surplus (the sum of producer and consumer surplus) from the trade fragmentation of each commodity. Such measure simultaneously accounts for the price and quantity effects of fragmentation, allowing us to identify commodities whose fragmentation may pose the highest economic risk at the country level.

Building on the new dataset and the model, we derive four main results. First, we document key features of commodity markets that make commodities more sensitive in the event of trade disruptions. As essential inputs to production and consumption, commodities are widely used across countries and exhibit low demand elasticities. Yet, commodity production is geographically concentrated, given the high concentration of natural endowments. For example, on average, the three largest suppliers of minerals account for about 75 percent of global production. As a result, commodities are highly traded, with many importers relying on only a few suppliers.

Second, further geoeconomic fragmentation could cause wide price differentials across blocs and lead to large commodity price changes. The scale of the price effects depends on the supply and demand imbalances caused by fragmentation and the price elasticities of supply and demand of individual commodities. In our baseline simulations, the price effects are particularly strong for some minerals critical for the green transition, such as copper, nickel, cobalt and lithium, and some highly traded agricultural goods such as soybean. This is due to the high concentration and uneven distribution of these commodities across the hypothetical trading blocs in our illustrative scenario.

Third, geoeconomic fragmentation of commodity markets would also lead to higher price volatility. Smaller markets in a fragmented world would provide fewer buffers against commodity-specific supply and demand shocks, leading to larger price responses than under free trade. Moreover, commodity producers would have powerful incentives to switch allegiances given potentially significant differences in commodity prices between blocs. This could induce more supply shocks, volatility, and uncertainty in commodity markets.

Finally, the impact of fragmentation on economic surplus depends not only on the price vulnerability but also on the value of production and consumption of the individual commodity relative to aggregate output. For example, energy commodity prices are not particularly vulnerable in the event of fragmentation in our baseline simulation, but the associated declines in economic surplus are more significant compared to the changes in surplus associated with minerals and some agricultural commodities, because these commodities are widely consumed and produced. Overall, there is significant heterogeneity in the potential surplus effects across countries. Due to vastly different and often offsetting impacts across net commodity-producing and net commodity-consuming countries, however, surplus losses appear relatively modest at the global level.

There are two main caveats to our results. First, the assumption of complete commodity trade fragmentation across blocs and free trade within blocs is highly stylized. In this respect, our model does not account for more realistic intermediate scenarios, such as partial trade between blocs, neutral countries, and arbitrageurs that would mute effects. Second, due to its partial equilibrium nature, our approach does not allow for the simultaneous disruption of trade of many commodities, which could have opposing or reinforcing effects within the same country. Because of these limitations, our results should not be interpreted as projected impacts from realistically complex geopolitical fragmentation scenarios. Rather, estimates should be viewed as commodity-specific measures of relative price-sensitivity and macro-economic fallout potential from trade fragmentation.

Our paper builds on several strands of literature. First, trade disruptions caused by the war in Ukraine and the COVID-19 pandemic have motivated numerous studies to gauge the macroeconomic effects of sectoral supply shocks, including those associated with commodity price movements (see Bonadio et al., 2021; Carvalho et al., 2021; Albrizio et al., 2023; Ilzetzki, 2022; Lafrogne-Joussier et al., 2022, among others). Our paper contributes by examining the economic impact of trade disruptions for a broad set of commodities.

Second, we contribute to a fast-developing literature that examines the consequences of geoeconomic fragmentation. Recent studies that quantify aggregate losses in this context include Attinasi et al. (2023), Felbermayr et al. (2023), Javorcik et al. (2022) and others.⁴ Most related and complementary to our analysis are Fally and Sayre (2018) and Bolhuis et al. (2023), who show that the specific features of commodities amplify the aggregate effects of trade disruptions in multicountry, multisector trade models. Bachmann et al. (2022), Chepeliev et al. (2022), Di Bella et al. (2022), Albrizio et al. (2023) and others focus on the implications of supply disruptions to natural gas and oil markets mostly based on second order approximations derived from multicountry multisector trade models.

Third, the paper adds to the literature examining drivers of commodity price fluctuations and their economic effects.⁵ We provide a simple model of trade disruptions as a source of shocks to supply and demand of commodities that affect prices and the economy for a large set of commodities.

Finally, our paper builds on previous studies by governments, international agencies and academics that document the concentration of production and trade as well as the widespread use of commodities with the goal of identifying those that are critical.⁶ Our paper extends this work by providing a transparent framework to assess commodities’ economic relevance and vulnerabilities in trade fragmentation scenarios.

The remainder of the paper is structured as follows. Section 2 describes the new dataset. Section 3 provides key stylized facts on the production and trade of commodities. Section 4 lays out a simple model of trade fragmentation in each commodity market, while Section 5 describes its calibration. Sections 6 and 7 present our results on commodities’ price changes and price volatility, respectively. Section 8 shows results on surplus changes. Section 9 discusses the sensitivity of the findings to alternative assumptions. Finally, Section 10 concludes and discusses avenues for future research.

2 A New Dataset

We assemble a new annual dataset on bilateral trade flows and production data at the country and commodity level. We focus on 48 energy, mineral, and agricultural commodities, as listed in Appendix Table A.6. They were selected because they represent a large share of global trade, or are part of critical raw materials lists of the EU or US. Commodities with insufficient data were eliminated from consideration (e.g., uranium).

Starting from the methodology of Fally and Sayre (2018) and updated by Bolhuis et al. (2023), we create the new dataset with several key innovations. First, we develop a set of adjustment factors to correct for different unit measurements for mineral commodities in the production and trade data, which is often overlooked in the trade literature. For instance, some minerals are expressed in metric tons of metal content in the production database, while their counterparts are presented in gross metric tons in the trade database. The factors convert the quantities in the trade data into equivalent metric tons of metal content (see Appendix Table A.1). These adjustment factors can be commodity- and country-specific.

Second, our approach includes the markets for mined upstream commodities (e.g. copper ore) and refined commodities (e.g. refined copper). Distinguishing between the different products along the value chain can lead to distinctly different production concentrations and trading patterns. The production and trade data for refined commodities also include recycled materials with the exception of aluminium.

Third, taking the concordances between minerals trade and production data from Fally and Sayre (2018) and Bolhuis et al. (2023) as a starting point, we develop concordances between the HS codes and commodity definitions used in production data. We do so based on consultations with the British Geological Survey (BGS) as well as the commodity-specific industry literature (see Tables A.2, A.3, and A.4 in the Appendix for the mapping between trade and production data for each commodity). For agricultural and energy commodities, we rely on concordances from the Food and Agriculture Organization (FAO) and the International Energy Agency (IEA), respectively.

Data on agricultural commodities and energy production and trade are from the FAO and the IEA. For minerals, production data come from the BGS and US Geological Survey (USGS), while trade data are from the Bilateral Commodity Trade Database (BACI), based on UN Comtrade. The data are for 2019, prior to the COVID-2019 pandemic.

3 Stylized Facts

Using the newly assembled dataset, we present several features of commodity markets that potentially raise the economic costs of disrupting their trade, despite commodities’ homogeneity and fungibility. These include the concentration of their production, the low elasticities of supply and demand, the high share of production that is traded as well as the high trade dependencies of importing countries.⁷

3.1 Concentration of Production

Unlike manufactured goods, the first stage of production of commodities depends on natural endowments that can be concentrated geographically. For instance, the extraction of energy and other mineral commodities requires cost effective geological deposits. The availability of fertile soil, water and adequate climate can constrain agricultural production and impact agricultural yields.

As a result of natural endowments’ concentration, there is significant concentration of production at the country level, with the top three producers accounting on average for about 65 percent of the global output of agricultural, about 50 percent of energy and about 75 percent of minerals commodities (Figure 2).⁸

View Full Size

Share of Global Production of the Top 3 Producers

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: For each commodity in the sample, the figure shows the share of global production that the top three producers account for in 2019.

• Download Figure
• Download figure as PowerPoint slide

For minerals, production is concentrated both at the mining stage due to the geographic concentration of deposits, but often also at the processing stage, with a few countries (most notably China), having developed a strong comparative advantage through the deployment of capital intensive facilities, efficient technological solutions, lighter environmental regulations, and relatively cheaper labor (see Figure B.1 in the Appendix).

3.3 Trade Dependence

With production highly concentrated and demand often spread across many countries, commodities are heavily traded. Their homogeneity and fungibility also contribute to market integration. As a consequence, the share of production that is traded internationally across commodity types is consistently higher than the ratio of world trade to gross output. Across agricultural commodities, around 40 percent of output is dedicated to trade, around 30 percent for energy and almost 50 percent for mineral commodities. For individual commodities, the share of traded output can be substantially higher: for example, more than 80 percent of lithium, or potash produced cross borders, with the share increasing over time for most raw commodities (see Figure B.3 in the Appendix).⁹

A corollary to the high concentration of commodity production is that a handful of key producer countries typically dominate the export side (see Figure B.4, panel (a) in the Appendix). As a result, most countries have not diversified their commodity imports. For instance, more than 60 percent of countries rely on less than three suppliers to satisfy their imports of key minerals such as cobalt, silver or nickel (Figure 3).

View Full Size

Import Dependency

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: For each commodity, the figure shows the share of countries that import a commodity from only one, two or three countries.

• Download Figure
• Download figure as PowerPoint slide

4 A Multi-Country Partial Equilibrium Commodity Market Model

To assess the economic impact of fragmentation, we develop a single-commodity model with multiple countries that face country-specific supply and demand curves:

where c denotes the country, $q_c^{s}and{q}_c^d$ are quantities supplied and demanded, p_c is the price, and η^s > 0 and η^d < 0 are the price elasticities of supply and demand. For simplicity and due to a lack of consistent country-specific estimates for a broader set of commodities, we assume that all countries face the same elasticities but have unique demand and supply shifters ${\gamma }_c^{d}and{q}_c^s$.¹⁰

Countries can be classified into one of two hypothetical blocs B ∈ {H, ROW}. Aggregating all countries within each bloc B, we get the following bloc-level demand and supply curves

where $Q_B^s={\Sigma }_{c\in B}{q}_c^s,Q_B^d={\Sigma }_{c\in B}{q}_c^d,{\gamma }_B^s=\ln \left({\Sigma }_{c\in B}{e}^{{\gamma }_c^s}\right)and{\gamma }_B^d=\ln \left(\Sigma {}_{c\in B}{e}^{{\gamma }_c^d}\right)$.

With this notation, we can define two market equilibria: one that allows trade between blocs and one that does not.

Integrated market equilibrium. The integrated market equilibrium must fulfill market clearing and non-arbitrage conditions such that $Q_H^s+Q_{ROW}^s=Q_H^d+Q_{ROW}^d$ and p_H = p_ROW = p_w, where p_w is the world price. The equilibrium world price can then be written as a function of supply and demand parameters. That is

where ${\Omega }^d\equiv \ln (e^{{\gamma }_H^d}+e^{{\gamma }_{ROW}^d})and{\Omega }^s\equiv \ln (e^{{\gamma }_H^s}+e^{{\gamma }_{ROW}^s})$. The equilibrium world price can be substituted into bloc- and country-specific demand and supply curves to obtain the corresponding quantities demanded, supplied, and net exports/imports.

Fragmented market equilibrium. The fragmented market equilibrium is defined as the one occurring from a full segmentation of the two blocs. Trade adjustments are assumed without cost across countries within a bloc. The equilibrium prices and quantities must then fulfill bloc-level market clearing conditions, inducing bloc-level prices. For bloc B, the fragmented market equilibrium price is given by

As before, the bloc- and country-level quantities and net exports are given by substituting bloc level prices, p_B, into the corresponding supply and demand curves.

Impact of fragmentation on commodity prices. In the model, the impact from market fragmentation is defined as the difference in country- and bloc-level quantities and prices between the integrated and fragmented market equilibria. The change in price is given by:

In our calibration, we standardize the initial world price to 1 in the model, so that

From initial equilibrium conditions, ${\Omega }^d=\ln (e^{{\gamma }_H^d}+e^{{\gamma }_{ROW}^d})=\ln (q_H^d+q_{ROW}^d)={\Omega }^s$ so Ω^d — Ω^s = 0. The change induced by fragmentation on bloc-level prices is therefore given by:

with changes in the quantities demanded and supplied obtainable by substituting (8) in the bloc-level demand and supply curves.

Equation (8) indicates that changes in prices (and quantities) depend on two factors. The first is the initial bloc-level trade imbalance $({\gamma }_B^d-{\gamma }_B^s)$, with an initially exporting (importing) bloc experiencing a drop (rise) in the bloc-level price (see also Figure 4a). This induces a price gap between blocs ex-post. The second factor is the elasticities of demand and supply. The more inelastic supply and demand are, the greater the price response of the quantities supplied and demanded (see Figure 4b).

View Full Size

A Single-Commodity Model of Fragmentation

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

• Download Figure
• Download figure as PowerPoint slide

Consumer, producer and total economic surplus. The simple model can be used to analyze consumer, producer and total surplus. These are calculated as the areas under the demand curve (above equilibrium prices), for consumer surplus changes, and above the supply curve (under equilibrium prices), for producer surplus changes. Total surplus is defined as the sum of consumer and producer surplus. Specifically, changes in consumer and producer surplus for country c are given by:

where p_c = p_B is the price in country c (which equals the bloc-level price in the fragmented equilibrium), and $p_w{q}_{c,w}^{d}and{p}_w{q}_{c,w}^s$ are the quantities, in US-dollar terms, demanded and supplied in the integrated market equilibrium. The changes in both producer and consumer surplus depend on the price ratio between the integrated and fragmented equilibria $\left(\frac{p_c}{p_w}\right)$, scaled by the supply and demand elasticity, and the total quantity demanded and supplied in dollar terms. Surplus changes are greater in economies that experience larger price changes in a more broadly consumed or produced commodity. This interaction is crucial to identify commodities with larger macroeconomic impact.

In the model, bloc-level total surplus necessarily falls with fragmentation, as shown in Appendix D. However, country-specific surplus need not decline. An exporter in an ex-ante importing bloc or an importer in an ex-ante exporting bloc will face more favorable prices that raise total surplus.

5 Calibration

We apply the model to a hypothetical baseline fragmentation scenario. We start by presenting the baseline allocation of countries across blocs. We then discuss the model calibration.

5.1 Baseline Bloc Configuration

Fragmentation in the baseline exercise is modelled along two hypothetical geopolitical blocs based on the March 2nd, 2022 United Nations (UN) vote on Russia’s war in Ukraine (see Table A.5 in the Appendix). For simplicity of exposition, we call the hypothetical bloc containing the US, and most European economies the “US-Europe+ bloc” and the bloc containing China and Russia, the “China-Russia+ bloc”.¹¹

The two-bloc scenario is meant to provide a clearly defined baseline and to make the exercise comparable to the recent literature. Countries’ geopolitical alignment could be partly driven by trade linkages and risk management strategies to reduce the fallout from spikes in geopolitical tensions. However, the endogenous formation of blocs remains beyond the scope of the paper. We examine alternative scenarios, including the impact of countries’ switching blocs and alternative bloc configurations further below.

Figure 5 and Figure B.5 in the Appendix provide some stylized facts about trade flows and production concentration across blocs. Indeed, as the production of some commodities is heavily concentrated in one of the hypothetical blocs, cross-bloc trade is substantial with integrated markets. Figure B.4, panel b, shows that China accounts for the largest share of imports of many commodities, particularly minerals.

View Full Size

Trade Flows Across Blocs (Net)

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: The diagrams present the value of trade across blocs in 2019, aggregated across countries and commodities within each commodity class. The red lines depict the baseline fragmentation scenario of no commodity trade between blocs.

• Download Figure
• Download figure as PowerPoint slide

The exercise assumes that, under fragmentation, commodities cannot be traded between blocs (as depicted by the red lines in Figure 5), while intra-bloc trade costs remain unchanged. Thus, all countries in the bloc face the same commodity price. Less severe scenarios, for example allowing partial interaction between blocs, would result in more muted effects, but those remain beyond the scope of the paper.

6 Fragmentation and Commodity Prices

Using the model, we compute the fragmentation-induced price changes for each commodity in the two blocs and rank the commodities.

The simulated price changes are directly proportional to the bloc-level commodity demand and supply imbalances triggered by the disruption of trade between blocs. Figure 6 presents the distribution of commodity price changes in the two blocs, while Figures 7a and 7b show the bloc-level simulated price changes for each commodity in the sample.

View Full Size

Commodity Price Changes Due to Fragmentation in Individual Commodity Markets: Distribution across Blocs and Commodities in the Baseline Scenario (Percent).

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: Price effects are capped at 500 percent for readability. “Energy” refers to coal, natural gas, and crude oil. The black squares in the bars represent the median, the bars the interquartile range, and the whiskers the data points within 1.5 times the interquartile range from the 25th or 75th percentile across commodities in the group. The dots indicate outliers. Selected commodities which experience price increases higher than 500 percent are labeled.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Commodity Price Changes Due to Fragmentation in Individual Commodity Markets in the Baseline Scenario (Percent).

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: Each bar represents the bloc-level commodity price change from fragmenting commodity trade. Changes in prices are capped at 150 percent for ease of exposition.

• Download Figure
• Download figure as PowerPoint slide

In our hypothetical baseline simulation, minerals tend to be the most vulnerable in the event of fragmentation, as their production is highly concentrated and unbalanced across blocs. The price of minerals at the mining stage would rise significantly in the China-Russia+ bloc, especially the prices of some of the key energy transition minerals (e.g. cobalt, lithium, copper and nickel as shown in Figure 7b). The production of these metals is concentrated in a handful of countries (Figure 2), while being largely consumed in the China-Russia+ bloc. At the same time, in the US-Europe+ bloc, prices of refined minerals would experience similar increases, driven by the prices of magnesium, platinum, palladium and aluminium. These commodities are mostly processed in China, South Africa and Russia (see Appendix Figure B.5).¹⁴

In contrast, the price effects of fragmentation on energy (crude oil, natural gas, and coal) and most agricultural (e.g., wheat, cotton, rice, maize) commodities are more muted due to the relative self-sufficiency of each bloc. Important exceptions are palm oil and soybeans, which could experience large price increases in the China-Russia+ bloc.¹⁵ Around 80 percent of the production of these broadly consumed commodities is concentrated in up to three countries (Indonesia and Malaysia for palm oil, the United States, Brazil and Argentina for soybeans), which are all part of the hypothetical US-Europe+ bloc in our baseline.

7 Fragmentation and Commodity Price Volatility

Geoeconomic fragmentation could not only affect the price level, but could also lead to higher price volatility. Jacks et al. (2011) show empirically that commodity price volatility is typically higher when markets are disintegrated. Based on our model, there are two channels that could be consistent with this empirical finding. First, smaller market sizes due to fragmentation result in a larger impact of any demand and supply shocks on prices. Second, the switching of blocs by countries represents an additional source of supply and demand shocks.

7.1 Smaller Market Sizes

In a fragmented world, markets become smaller and bloc-level prices would be more responsive to country-level shocks (see also Albrizio et al., 2023). In the partial equilibrium single-commodity model, the price response is proportional to the supply shock’s size relative to the overall market. Thus, by restricting the set of countries they trade with, countries would face larger price increases in response to the same negative supply shock.

Given the expression for ln(p_B) in equation (7), it is straightforward to show that the elasticity of the price of a commodity to country-level supply shocks is

$$\begin{array}{cc}\frac{\partial \ln p}{\partial {\gamma }_c^s}=-\frac{1}{{\eta }^s-{\eta }^d}\frac{e^{{\gamma }_c^s}}{{\Sigma }_c{e}^{{\gamma }_c^s}},& \left(11\right)\end{array}$$

with the second term representing the production share of country c in a given commodity. For a given numerator, the denominator of the second ratio would be smaller in a fragmented world, leading to higher price responsiveness to a given country-level supply shock.

In an illustrative example, Figure 8 compares the impact on wheat prices of a three-standard deviation shock to the US wheat production in an integrated market with that of a fragmented market. The same supply shock has double the impact on wheat prices when the market is fragmented into two smaller blocs.¹⁶ This is important, as climate change is expected to raise the variability of agricultural output. Overall, in fragmented markets, the price response to supply and demand shocks is amplified.

View Full Size

Wheat Price Increases in the US-Europe+ bloc from a US Wheat Harvesting Shock in an Integrated versus Fragmented World (Percent).

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: The bars in the figure depict the wheat price increase in the US-Europe+ bloc in response to a three-standard-deviation negative shock to US wheat production. The figure compares the price increases in a free-trade world to those in a fragmented world.

• Download Figure
• Download figure as PowerPoint slide

7.2 Countries Switching Blocs

In a fragmented world, major commodity producers would face powerful incentives to switch geopolitical allegiances, representing a new source of supply and demand shocks. For highly concentrated commodity markets, a single exporting country switching to the other bloc can lead to a large supply gap and trigger substantial price changes. Uncertainty about a country’s geopolitical alignment could lead to price volatility itself, as traders update their priors regarding potentially large fragmentation-induced price swings.

7.2.1 Maximum Commodity Price Increases from a Single Exporting Country Switching Blocs

Starting from the fragmented equilibrium, the change in prices from one country leaving (joining) the bloc is given by:

$$\begin{array}{cc}\ln \left(p_{B^{\prime }}\right)-\ln \left(p_B\right)=\frac{{\gamma }_{B^{\prime }}^d-{\gamma }_{B^{\prime }}^s}{{\eta }^s-{\eta }^d}-\frac{{\gamma }_B^d-{\gamma }_B^s}{{\eta }^s-{\eta }^d},& \left(12\right)\end{array}$$

where B’ is the new bloc without (with) the leaving (joining) country. For a country c leaving, using the aggregating equation and calibration above, we get

$$\begin{array}{cc}{\gamma }_{B^{\prime }}^d=\ln (Q_{B,w}^d-q_{c,w}^d)& \left(13\right)\end{array}$$

$$\begin{array}{cc}{\gamma }_{B^{\prime }}^s=\ln (Q_{B,w}^s-q_{c,w}^s)& \left(14\right)\end{array}$$

Figure 9 shows the distribution of the greatest price increase in a bloc that can be induced by a single exporter switching alliances in each commodity market. The underlying top 15 most vulnerable commodities are presented in Figures 10a and 10b.

View Full Size

Largest Price Increases Induced by a Single Exporter Switching Blocs (Percent)

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: Price changes are capped at 800 percent for readability. Each observation in the box plots represents the largest price increase that a commodity can experience in each bloc from a single exporting country’s switching to the other bloc. Note that the US (China) are not allowed to switch away from the US-Europe+ (China-Russia+) bloc. The black squares in the bars represent the median, the bars the interquartile range, and the whiskers the data points within 1.5 times the interquartile range from the 25th or 75th percentile across commodities in the group. The dots indicate outliers; commodities representing the largest outliers are labeled.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Top 15 Largest Price Increases from a Single Exporter Switching Blocs (Percent)

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: Each bar represents the largest bloc-level price increase that the corresponding commodity experiences from a single exporting country switching to the other bloc.

• Download Figure
• Download figure as PowerPoint slide

Given their highly concentrated production, minerals at the mining stage tend to be most sensitive to exporters’ switching blocs. For example, South Africa produces one-third of the world’s manganese, a metal used in steel-making and batteries. If South Africa switched to the hypothetical US-Europe+ bloc, the price of manganese in the China-Russia+ bloc could rise more than 800 percent.

7.2.2 Accounting for Countries’ Probability to Switch Blocs

Not all countries are equally likely to switch allegiances in a fragmented world. A commodity market may experience higher volatility if the pivotal country behind the maximum price effect discussed in the previous section is also more likely to switch. Conversely, the price changes shown above might be less relevant if the countries behind those effects are core members of a hypothetical geopolitical bloc. This section constructs a measure that accounts for the likelihood that countries change their allegiances.

We first test how well each country’s assigned bloc can be predicted by its economic and military proximity to the “core” countries in our hypothetical blocs.¹⁷ Specifically, consider the model

$$b_c=\log it({\beta }_0+\underset{j\in \{US,EU,RUS,CHN\}}{\Sigma }[{\beta }_E^j{E}_c^j+{\beta }_M^j{M}_c^j]+{\epsilon }_c)$$

where b_c is a dummy that equals 1 when country c is in the US-Europe+ bloc, and 0 when it is in the China-Russia+ bloc. Economic distance is measured by the trade flows between country c and each of the core members of each bloc; specifically, $E_c^j=1-\frac{s_{x,c}^j+s_{i,c}^j}{2}$, where $s_{x,c}^j$ is the share of exports from country c that are destined to country j, and $s_{i,c}^j$ is the share of imports by country c that come from country j. Military distance is measured as $M_c^j=1-{\sigma }_c^j$, where ${\sigma }_c^j$ is a similarity score between the portfolio of military alliances of countries c and j (Signorino and Ritter, 1999), measured using data from the Alliance Treaty Obligations and Provisions (ATOP) project (see Leeds et al., 2002).¹⁸ For better interpretation, the military distance measure is normalized so the standard deviation of its cross-country distribution is 1. We also include a constant term β₀ and the residual is ε_c.

Estimated coefficients using data from 2018 are shown in Table 1. Based on those coefficients, military distance appears to be a better predictor of bloc positioning than economic distance, with countries that are more distant from the US and Europe being less likely to be in the US-Europe+ bloc (similarly, countries more distant from Russia are more likely to be in the US-Europe+ bloc).¹⁹

^{Table 1:}

Probability of Being in the US-Europe+ Bloc

Robust standard errors in parenthesis; ***, **, and * indicate coefficients are statistically different from zero at the 1%, 5%, and 10% levels, respectively.

^{Table 1:}

Probability of Being in the US-Europe+ Bloc

Model	Coefficients
	${\beta }_E^{US}$	${\beta }_E^{EU}$	${\beta }_E^{CN}$	${\beta }_E^{RU}$	${\beta }_M^{US}$	${\beta }_M^{EU}$	${\beta }_M^{CN}$	${\beta }_M^{RU}$	N	R2
Logit	1.07 (3.20)	-1.50 (1.51)	3.43 (2.16)	13.59 (9.39)	-1.03*** (0.40)	-2.47** (1.04)	-1.59 (1.10)	3.82** (1.69)	156	0.22
Linear	0.23 (0.33)	-0.21 (0.22)	0.75** (0.36)	1.08* (0.58)	-0.15*** (0.04)	-0.26*** (0.09)	-0.37*** (0.13)	0.57*** (0.17)	156	0.24

Robust standard errors in parenthesis; ***, **, and * indicate coefficients are statistically different from zero at the 1%, 5%, and 10% levels, respectively.

View Table

^{Table 1:}

Probability of Being in the US-Europe+ Bloc

Model	Coefficients
	${\beta }_E^{US}$	${\beta }_E^{EU}$	${\beta }_E^{CN}$	${\beta }_E^{RU}$	${\beta }_M^{US}$	${\beta }_M^{EU}$	${\beta }_M^{CN}$	${\beta }_M^{RU}$	N	R2
Logit	1.07 (3.20)	-1.50 (1.51)	3.43 (2.16)	13.59 (9.39)	-1.03*** (0.40)	-2.47** (1.04)	-1.59 (1.10)	3.82** (1.69)	156	0.22
Linear	0.23 (0.33)	-0.21 (0.22)	0.75** (0.36)	1.08* (0.58)	-0.15*** (0.04)	-0.26*** (0.09)	-0.37*** (0.13)	0.57*** (0.17)	156	0.24

Robust standard errors in parenthesis; ***, **, and * indicate coefficients are statistically different from zero at the 1%, 5%, and 10% levels, respectively.

We use the logit model above to calculate the predicted probabilities that each country belongs to the US-Europe+ bloc, ${\hat{p}}_c$. Since b_c is simply a dummy variable, we set the probability that any given country will permanently switch from their current bloc as

$$\mathbb{P}\left({\text{switch}}_c\right)=|b_c-{\hat{p}}_c|.$$

Thus, if a country is in the US-Europe+ bloc, its probability of switching is $1{-\hat{p}}_c$; if the same country is instead placed in the China-Russia+ bloc, then its probability of switching is simply ${\hat{p}}_c$. In either case, the probability of switching is equal to the model-implied likelihood that the country is placed in the wrong bloc. Intuitively, a country is less likely to switch from their current bloc when assigned to the bloc that contains countries to which they are historically closer to, in both economic and military terms.

This exercise allows us to weigh the price change induced by a single exporter switching by the probability of each country switching, as illustrated in Figure 11. Comparing this figure, to Figures 10a and 10b we find that the top most vulnerable commodities remain broadly unchanged in both groups once we account for differences in switching probabilities. This implies that major exporters switching blocs and the associated supply shocks would be a real concern in a fragmented world.

View Full Size

Top 15 Expected Price Increases from Exporting Countries Switching Blocs Weighted by the Probabilities of Switching (Percent).

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: Each bar represents the expected average bloc-level price increase that the corresponding commodity experiences from individual exporting countries switching to the other bloc weighted by the probability of each exporting country switching.

• Download Figure
• Download figure as PowerPoint slide

8 Fragmentation and Economic Surplus

We use changes in consumer, producer and total surplus due to the disruption of trade in commodities to measure individual commodities’ macro-relevance. This measure accounts for both changes in price and the importance of each commodity in overall consumption and production.

Several findings are revealed by the analysis, as shown in Figure 12a. First, inefficiencies associated with restricting free trade result in losses in bloc-level surplus, and consequently, the global economy is worse of from fragmentation of trade in individual commodities. This finding is independent of the precise definition of blocs as demonstrated in Appendix 9.2 and in Appendix D . Global economic losses, as captured in surplus changes, are, however, small in magnitude.

View Full Size

Changes in Surplus due to Fragmentation in Individual Commodity Markets

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: In panel 1, each data point in the box plots represents the total bloc-level surplus change from fragmenting trade in a single commodity. The black squares in the bars represent the median, the bars are the interquartile range, and the whiskers reflect the data points within 1.5 times the interquartile range from the 25th or 75th percentile across commodities in the group. Dots indicate outliers; the commodities associated with the largest surplus declines are labeled. Data labels use International Organization for Standardization (ISO) country codes. GNE = gross national expenditure.

• Download Figure
• Download figure as PowerPoint slide

Second, bloc- and country-level changes in total surplus are generally small as a share of Gross National Expenditure (GNE) with notable exceptions. Figure B.8b in the Appendix show that fragmentation in palm oil and copper markets could cause declines in surplus of over 1 percent of GNE in the China-Russia+ bloc.

Two reasons underpin the mostly moderate effects: (i) changes in total surplus reflect changes in consumer and producer surplus, which move in opposite directions within a country or a bloc, for a given price change. When the price of a commodity increases, part of the decline in consumer surplus is offset by an increase in producer surplus; and (ii) many of the most price-sensitive commodities do not represent large shares in consumption or production.

Restricting trade in commodities that are less price-vulnerable in the event of fragmentation could still lead to sizable surplus declines. For example, energy commodities are not particularly price vulnerable under the baseline bloc configuration, but the associated declines in surplus are more significant. This is because energy commodities are widely consumed and produced, and even small price changes matter. In contrast, the changes in surplus associated with disruptions in the more-price-sensitive minerals’ trade are more subdued due to their limited relevance in countries’ production and consumption.²⁰

Importantly, bloc-level aggregation masks important heterogeneities across countries (see Figure B.9 in the Appendix). Within each bloc, some countries experience an increase in surplus (net-exporting countries in a net-importing bloc, and net-importing countries in a net-exporting bloc) and some experience a decline, as shown in Figure 13 and in the theoretical proof in Appendix D. Such changes can be quite sizable for a few commodity importers and exporters.²¹ For instance, Figure 12b plots the surplus changes in the top two net exporters of copper, oil and palm oil across the two blocs. Fragmentation of trade in copper ores and concentrates would reduce surplus by 2.5–5 percent of GNE in Peru and Chile, the largest copper exporters in the US-Europe+ bloc, where copper prices would fall. At the same time, fragmentation would lead to large surplus gains in Mongolia and Kazakhstan, who would substantially scale up production and export at higher prices in the copper-scarce China-Russia+ bloc.

View Full Size

Distribution of Surplus Changes Across Exporting and Importing Countries in the Two Blocs.

Citation: IMF Working Papers 2023, 201; 10.5089/9798400252426.001.A001

Note: The figure plots the distribution of changes in total surplus due to fragmentation in individual commodity markets as a share of gross national expenditure. Each observation in the distribution is a commodity-country combination. Separate histograms are plotted for two groups of countries: (i) net-commodity-exporting countries in a commodity-importing bloc and net-commodity-importing countries in a commodity-exporting bloc [blue], and (ii) net-commodity-exporting countries in commodity-exporting blocs, and net-commodity-importing countries in commodity-importing blocs [red]. Countries’ importing/exporting status is based on the pre-fragmentation baseline.

• Download Figure
• Download figure as PowerPoint slide

9 Robustness Analysis

The results from our baseline simulation are robust to alternative price elasticities of supply and demand. In line with the sensitivity of the simulated price changes to major producing countries switching blocs, the precise bloc configuration is an important factor underpinning the results for individual commodities. In highly concentrated markets, the assignment of major producers across blocs often makes a difference.

10 Conclusion

This paper examined the economic implications of trade fragmentation in individual commodities, using a novel dataset of production and bilateral trade flows of the 48 most important energy, mineral and agricultural commodities. Commodities appear highly vulnerable in the event of fragmentation due to their highly concentrated and difficult-to-relocate production, hard-to-substitute consumption, and their critical role as inputs for manufacturing and key technologies.

Our simple partial equilibrium framework suggests that fragmentation could cause wide price differentials across blocs and lead to large changes in commodity prices, depending on the resulting supply and demand imbalances and commodities’ elasticities of supply and demand. In our illustrative simulation, some minerals critical to the clean energy transition and some highly traded agricultural goods are among the most vulnerable in the event of fragmentation, implying risks to global climate change goals and food security. Moreover, potential economic impacts from fragmentation differ vastly across countries, with offsetting effects across consumer and producer countries resulting in modest surplus losses at the global level.

Our results also show that a fragmented world would be more volatile. Commodity price volatility could intensify due to smaller market sizes and the incentives for producers to switch geopolitical allegiance. This could result in more volatile inflation, making monetary stability more challenging.

Comprehensively mapping vulnerable commodity markets and fragmentation implications across multiple global markets required several simplifying assumptions, which merit refinement – especially in those markets identified as most vulnerable and macroeconomically impactful in the event of fragmentation. Further research could more closely examine the impact of fragmentation of critical minerals on the clean energy transition. A deeper understanding on the intra-country effects of fragmentation on food security, in particular in low-income countries would be another highly policy relevant research topic. The role of recycling in dampening the effects of fragmentation on minerals markets could be explored further. Finally, a consistently estimated set of price elasticities of demand and supply could provide important foundations for a more rigorous calibration of models with disaggregated commodity markets.