## Section 6 concludes. ## Source details **Canonical URL:** [Section 6 concludes.](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025178-source-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025178-source-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025178-source-pdf.pdf.json) --- ### Updated and expanded dataset — scope, collection, and harmonization - Dataset: unbalanced panel of invoicing currency shares in imports and exports of 132 countries from 1990 to 2023; invoicing currency shares generally refer to goods trade. - Country outreach and data collection: - Contacted national authorities of around 120 countries between June 2024 and July 2025. - Primary contacts: governors’ offices and senior officials in statistics, payments and international departments of central banks; subsequently statistics offices, ministries of finance, customs/revenue authorities when needed. - For more than half of the countries in the dataset, data were obtained by asking national authorities to compile information from raw customs records. - EU data sources: - Annual ECB IRE data collection exercise (as in Boz et al. (2022)). - Non-public ECB archive data not reported in recent IRE exercises (as in Boz et al. (2022)). - Eurostat data (as in Boz et al. (2022)). - New: direct requests to national central banks (via the Eurosystem’s Working Group of External Statistics) to obtain invoicing currency information—especially for the renminbi (CNY). - Non-EU approach: systematic online search for official data; where not publicly available, direct requests to national authorities leveraging contacts across IMF, ECB, EBRD, AfDB, AACB, ADB, BIS, SEACEN. - Harmonization challenges: large cross-country heterogeneity in definitions, coverage (currencies, goods vs. services), reporting procedures, and whether customs forms record invoicing vs. settlement currency. ### Data properties, imputation, coverage, and limitations - No global harmonized standard existed for reporting invoicing currency until the IMF BPM seventh edition published in March 2025; the IMF BPM encourages reporting invoicing currency as a supplementary item but only from 2029-30 onward. - Proxying: in some cases settlement currency is used as a proxy for invoicing currency when invoicing currency is unavailable. - Balanced-panel construction (for aggregate trend analysis): - Linearly interpolate between missing observations. - Backpolate before the first available observation by assuming earlier observations had the same value as the earliest available. - Extrapolate after the last available observation by assuming later observations had the same value as the latest available. - Note: this simple extrapolation procedure stacks the deck against finding secular trends (e.g., rise in renminbi, decline in dollar invoicing). - After interpolation, backpolation and extrapolation, the dataset covers around 75% of global exports. - Maximum country coverage for dollar export invoicing in any year: 121 countries (in 2022). - EU countries account for roughly one-third of the share of global exports covered by the dataset. - Notable omissions and data gaps: - Customs authorities do not record invoicing currency or record it with insufficient accuracy; settlement currency proxy unavailable (examples: Hong Kong, Singapore, Mexico). - Unable to establish contact with relevant authorities (examples: Nigeria, Vietnam). - Only historical observations available from Boz et al. (2022) with no updates (examples: Ghana, Malawi, Tanzania). - Limited or single historical observations due to data storage changes or lack of access (examples: Canada — single observation for 2001 from Kamps (2006); India — no information after 2014 due to technical infrastructure changes; Pakistan — confined to 2001-2003 data from Kamps (2006)). - Geographic gaps: relatively limited information for Central America and Sub-Saharan Africa. - Comparative coverage versus Boz et al. (2022): - Information on US dollar and euro invoicing patterns for about 20% more countries compared to Boz et al. (2022). - Additional information for 2020-23 for generally all countries. - Renminbi invoicing patterns: information for more than 100 countries (substantially expanded relative to Boz et al. (2022)). - Note on interpretation: total dataset includes 132 countries but export/import country counts by currency may be lower because some countries report only imports or only exports for a currency. ### China settlement currency and its (lack of) validity as a proxy for invoicing - Context: PBoC measures to promote the renminbi in trade settlement (pilot program 2009, expanded 2012; offshore markets; swap agreements; Shanghai Petroleum and Natural Gas Exchange launched 2018; Belt and Road Initiative emphasis). - SAFE settlement data (China): share of China’s exports settled in renminbi rose to 50% in 2023; US dollar share declined from 90% to 50% (SAFE data). - Back-of-the-envelope 2023 plausibility check using updated dataset: - About 1.5% of total imports of the countries in the dataset are invoiced in renminbi = USD 216 billion. - If China’s trade settlement currency corresponded to trade invoicing currency, then USD 1,646 billion of China’s exports would be invoiced in renminbi. - This implies USD 1,430 billion of imports of countries not in the dataset would need to be invoiced in renminbi. - Actual imports from China by countries not in the dataset = USD 1,373 billion → required USD 1,430 billion implies 104% of their total imports from China must be invoiced in renminbi. - As a share of their total imports from the world this would be 18%, versus 1.5% for total imports of countries in the dataset — a discrepancy judged implausible. - Alternative conservative implied-share calculation: - If countries not in dataset invoice the same share of imports in renminbi as countries in dataset (and renminbi invoicing occurs only in bilateral trade with China), implied share of China’s exports invoiced in renminbi = 6.5%, much lower than SAFE settlement share 50%. - Conclusion: - Given implausibly large adjustments required and large gap between implied invoicing shares and SAFE settlement shares, settlement currency (SAFE) is not used as a proxy for invoicing currency in China’s trade in this dataset. - Consistent with Boz et al. (2022), SAFE trade settlement currency data are not included as a proxy for trade invoicing currency for China. ### Aggregate patterns: USD and EUR invoicing - Evolution and roles: - Combined share of dollar and euro invoicing has remained fairly stable over time. - The dollar is the dominant vehicle currency in global trade: the share of exports invoiced in dollars far exceeds the share of exports to the US. - The share of global exports invoiced in euros is about as large as the share of global exports to the euro area, consistent with the euro being used mainly in trade with the euro area rather than as a vehicle currency in third-country trade. - No significant shifts in dollar and euro invoicing shares at the global level from 2020 to 2023—the four additional years in the dataset compared to Boz et al. (2022). - Robustness excluding issuer-area countries: - Excluding euro area countries and the US: - The share of the dollar in global trade becomes an even greater multiple of the share of global exports destined to the US. - The share of the euro in trade invoicing is almost halved and is—at best—about as large as non-euro area countries’ exports to the euro area. - Dropping euro area countries reveals global euro invoicing and exports to the euro area have been converging, with growth in euro invoicing traced to non-euro area European countries. - Pegged currencies treatment: - Invoicing in currencies pegged to the dollar or the euro is accounted for by adding invoicing in home currency to invoicing in dollar/euro for countries whose currencies are pegged to the dollar/euro. - Peg classification per Ilzetzki et al. (2019) fine assessment categories "1", "2", "3", "4"; classification over full sample if pegged for every year since 2005 until 2023. - Home currency added to dollar invoicing shares for: Bahrain, Brunei Darussalam, Curacao, Jordan, Macao, Maldives, Saudi Arabia and Sint Maarten; except for Saudi Arabia, the home currency shares are low. - Home currency added to euro invoicing shares for: Bosnia and Herzegovina, and the Republic of Congo. ### Renminbi invoicing trends and regional patterns - Renminbi patterns: - Share of global exports to China has increased strongly since 2000 and is now comparable to the share of exports to the US. - The share of global exports invoiced in renminbi is hardly visible on the same scale as the dollar and the euro. - Using raw (rather than inter/extrapolated) renminbi invoicing shares to capture recent growth: - Renminbi invoicing was hardly visible until 2010; since then it has increased across countries in most regions, particularly in Asia and more recently in Europe and Latin America. - Despite recent strong increases, renminbi invoicing remains below the share of global exports to China. - Figure-based robustness: zoomed conclusions similar when looking at unweighted averages across countries and when excluding euro area countries or Russia. ### Geopolitics and invoicing — empirical approach and sample shares - Geopolitical alignment measure: ideal point distance metric based on UN General Assembly voting patterns (Bailey et al. 2017). - Classification: upper (lower) 25% percentile in the distribution of the distance to the US in 2023 across countries = US most-aligned (US least-aligned); US neutral group = 50% of countries. - Country-group export shares: - 34% of world exports originate from US most-aligned countries. - 8% of world exports originate from US least-aligned countries. - 9% of world exports originate from the US itself. - 49% of world exports originate from US neutral countries. ### Key empirical findings on geopolitics and currency usage - Dollar: - US least-aligned countries invoice a larger share of their exports in dollars compared to US most-aligned countries, despite similar shares of exports to the US. - US least-aligned countries are predominantly emerging market and developing economies, which tend to invoice in dollars more frequently than advanced economies. - Dollar-invoiced trade has declined since the early 2010s in US least-aligned countries, while it has been stable in US most-aligned countries. - The long-term decline in dollar invoicing among US least-aligned countries has been driven primarily by Russia and Saudi Arabia. - Euro: - Share of exports invoiced in euros is substantially lower in euro area least-aligned countries than in euro area most-aligned countries. - Euro plays little or no role as a vehicle currency in euro area least-aligned countries. - Share of exports invoiced in euros in euro area most-aligned countries has been increasing, with growth traced to non-euro area European countries. - Renminbi and trade with China: - China has been gaining importance as an import source, especially for US least-aligned countries, with divergence after the COVID-19 trade drop in 2020 and Russia’s invasion of Ukraine in 2022. - Renminbi invoicing has "taken off" in US least-aligned countries; increases in US most-aligned countries have been modest and reversed since 2022. - Divergence is stark when Russia included in US least-aligned group, but remains visible when Russia is excluded. - Decompositions show renminbi invoicing growth among US most-aligned countries largely due to European countries; among US least-aligned countries mostly due to Asian countries. - Five summary takeaways: 1. US least-aligned countries tend to invoice a larger share of their exports in dollars compared to US most-aligned countries. 2. Because US least-aligned countries account for a smaller share of global trade, the value of their exports invoiced in dollars remains lower than that of US most-aligned countries. 3. The decline in dollar invoicing by US least-aligned countries since the early 2010s is not broad-based and has been driven primarily by Russia and Saudi Arabia. 4. Euro invoicing in global exports is predominantly by euro area most-aligned countries. 5. Trade with China and renminbi invoicing have been growing faster for US least-aligned countries since 2014, and at least since 2021 not only due to Russia. ### Country-level shifts, decompositions, and illustrative case of Russia - Decomposition of dollar export invoicing among US least-aligned countries: - Primary drivers of group-level decline: Russia and Saudi Arabia. - Figure 7: stacked and unstacked country-level contributions to group decline. - Renminbi invoicing and imports from China by alignment: - Figures show imports from China and imports invoiced in renminbi by country-group; vertical lines mark 2022. - Renminbi invoicing increases concentrated in US least-aligned countries. - Russia illustrative case: - Changes in renminbi and euro invoicing shares closely track shifts in bilateral trade shares. - Rise in renminbi invoicing in Russia primarily reflects reorientation of imports sourced from China being invoiced in renminbi. - Decline in euro invoicing corresponds to reduced euro-invoiced trade with the euro area. - For Russia, trade settlement currency is used as a proxy for invoicing currency; settlement and customs invoicing data show similar trends for USD, EUR, CNY, RUB. - Evidence of Central Bank of Russia leveraging swap lines with PBoC to provide renminbi liquidity for imports, affecting settlement vs invoicing distinctions. ### Changes in geopolitical distance and corresponding invoicing shifts (country-level evidence) - Continuous measure used: change in ideal point distance (Bailey et al. (2017)) comparing 2015-19 vs 2022-23. - Observed bilateral distance shifts: - Majority of countries moved closer to the US and away from China. - Small group (including Russia and Belarus) moved farther from the US and closer to China. - Among countries moving closer to the US and away from China, most also became more distant from euro area. - Only a few countries distanced themselves from both the US and the euro area (Russia, Belarus, Kazakhstan, Uzbekistan, Mali). - Dollar invoicing changes versus geopolitical distancing: - Countries with largest drops in dollar export invoicing shares in 2022-23 vs 2015-19 include Kyrgyzstan, Russia, Belarus, Uzbekistan. - Negative association between increases in geopolitical distance from the US and declines in dollar export invoicing shares across full sample. - Caveat: regression plots do not prove causality beyond compositional bilateral trade effects. ### Panel regression evidence — methodology and key quantitative results - Sample period: 1999-2023. - Main specification regresses exporter-currency export invoicing share s^c_{e,t} on: - lagged dependent variable s^c_{e,t-1}; - controls w^c_{e,t}: share of exporter’s exports to issuer of currency c, share of exports to economies that peg to that issuer, commodity/export-in-oil shares, bilateral nominal exchange rate between exporter’s currency and the dollar/euro/renminbi; - exporter and time fixed effects α^c_e and τ^c_t. - Regularization: exclude observations when first difference over time falls below 1st percentile or exceeds 99th percentile. - Inference: Driscoll-Kraay standard errors robust to serial correlation and cross-section dependence. - Selected summary statistics (Table 1): - USD invoicing share: mean 49.15; min 0.45; p5 5.96; p50 42.96; p95 98.61; max 100.00; sd 33.33; count 1,307 - EUR invoicing share: mean 41.24; min 0.00; p5 0.29; p50 39.44; p95 91.94; max 99.46; sd 35.26; count 1,303 - CNY invoicing share: mean 0.25; min 0.00; p5 0.00; p50 0.00; p95 1.17; max 29.32; sd 1.40; count 704 - Share of trade with US: mean 7.63; min 0.00; p5 0.45; p50 4.59; p95 27.73; max 44.73; sd 8.41; count 1,307 - Share of trade with euro area: mean 34.65; min 0.05; p5 3.04; p50 38.43; p95 67.59; max 80.68; sd 21.71; count 1,303 - Share of trade with China: mean 7.24; min 0.00; p5 0.21; p50 2.57; p95 26.52; max 92.60; sd 10.36; count 704 - Commodity trade share: mean 23.40; min 0.02; p5 2.87; p50 14.04; p95 74.80; max 98.58; sd 22.69; count 1,250 - Oil trade share: mean 6.38; min 0.00; p5 0.00; p50 2.22; p95 34.30; max 55.58; sd 10.18; count 1,307 - Bilateral exchange USD exchange rate (log): mean 440.52; min 111.37; p5 366.41; p50 449.91; p95 476.33; max 498.24; sd 39.29; count 1,307 - Main regression findings (Table 2): - Persistence: lagged invoicing share coefficients large and precisely estimated (values reported as 0.73, 0.75, 0.80, 0.80, 0.92, 1.02 with p-values (0.00)). - Bilateral trade share with issuer: positive coefficients (0.11, 0.06, 0.08, 0.09, 0.01, 0.01) indicating exporters invoice more in issuer currency the more they trade with issuer. - Commodity and oil trade: - Commodity trade share coefficients: 0.09, −0.04, −0.00 with p-values (0.00)(0.04)(0.74). - Oil trade share coefficients: 0.17, −0.16, −0.01 with p-values (0.01)(0.00)(0.04). - Interpretation: commodity/oil trade positively correlated with dollar invoicing and negatively with euro and renminbi invoicing. - Within R-squared reported across columns: 0.65, 0.65, 0.77, 0.78, 0.80, 0.96 (observations and country counts reported per column in Table 2). - Geopolitical distance effects (Table 3 and Table 4): - Baseline: over full sample, dollar and euro invoicing use not systematically related to geopolitical distance to issuers; renminbi use greater for exporters closer to China. - Since 2022 (post-2021 interactions): - Correlation between dollar/euro invoicing and geopolitical distance to issuers has become more negative. - For the euro the overall negative effect since 2022 is statistically significant at conventional levels. - Renminbi negative correlation with distance from China has become more negative since 2022 (significance depends on controls). - Table 4 patterns (selected reported numbers): - Geopolitical distance to the US × post-2021, oil-share control, column (8): dollar −0.36 (parenthetical p-value (0.09)). - Geopolitical distance to China × post-2021, column (8): 0.86 (parenthetical p-value (0.00)). - Summary: since 2021/2022 greater distance from the US associated with lower dollar use; distancing from US associated with greater use of renminbi, home, and other third currencies; distancing from the euro area reduces euro use and is associated with increases in renminbi, home, and other currencies; distancing from China (when controlling for oil share) associated with greater dollar invoicing. - Robustness: - Results similar when using import invoicing shares, when using trading-partner average geopolitical distance, and after excluding Russia/Belarus/Uzbekistan for some specifications. - Pooled exporter–invoicing currency regressions impose homogeneity and can yield different estimates; authors focus on currency-specific regressions. ### Commodity and oil invoicing — mechanisms and evidence - Theoretical channels: - Commodities, particularly oil, predominantly invoiced in dollars historically. - Homogeneous commodity pricing and organized exchange trading favor single-currency quoting. - Invoicing currency of intermediate inputs influences exporters via marginal costs channel. - Settlement currency restrictions (sanctions etc.) can reduce attractiveness of a currency for invoicing. - Anecdotal non-dollar settlement cases (reported sources): - Russia settled oil exports to China in renminbi (Reuters 2023c). - Iran, Venezuela reported to shift to renminbi for China settlements (CNN 2012, Reuters 2019). - Venezuela reportedly settling more oil in cryptocurrencies (Reuters 2024). - India settled oil imports from UAE in Indian rupees (Reuters 2023a) and some oil imports from Russia in renminbi; discussions of settling in rubles (Reuters 2023b; Bloomberg 2024). - Reports of negotiations for Saudi Arabia to settle portion of oil exports to China in renminbi (Wall Street Journal 2022). - Empirical suggestive analysis on oil: - Descriptive: weakening co-movement between world oil prices and oil export shares may indicate shift away from dollar invoicing (alternative explanations include price caps, compositional effects). - Russia: recent weakening of co-movement between world oil prices and Russia’s oil export share — consistent with a shift away from dollar invoicing but could reflect G7 price cap effects. - Canada (US-aligned benchmark) shows strong co-movement, consistent with dollar invoicing persistence. - Regression (Table 5) testing whether correlation between oil exports and dollar invoicing declined for US least-aligned exporters after 2021: - Oil trade share coefficients: - Column (1): 0.18 (standard error (0.00)) - Column (2): 0.25 (standard error (0.00)) - Column (3): 0.26 (standard error (0.00)) - Oil trade share × least-aligned × post-2021 interaction coefficients: - Column (1): −0.04 (standard error (0.30)) - Column (2): −0.02 (standard error (0.49)) - Column (3): 0.03 (standard error (0.45)) - Trade to GDP (columns (2)–(3)): −0.01 (standard errors (0.71), (0.46)) - Within R-squared: Column (1) 0.65; Column (2) 0.66; Column (3) 0.65 - Observations: Column (1) 1307; Column (2) 1283; Column (3) 1262 - Countries: Column (1) 111; Column (2) 107; Column (3) 105 - Column (3) excludes Belarus and Russia relative to column (2). - Interpretation: interaction estimates negative in some specifications but estimated imprecisely; authors conclude impossible to ascertain whether negative sign reflects true decline or data noise. - Robustness: commodity-based specifications (Table B.8) consistent but no decisive evidence of broad de-dollarization in non-oil commodities. ### Next analytical steps and authors’ planned regressions - Authors indicate intent to: - Use regression analysis to further explore role of commodity trade developments and geopolitical alignment in dollar invoicing decline. - Examine whether invoicing shifts have broadened beyond Russia and Saudi Arabia since Russia’s 2022 invasion and whether changes go beyond compositional bilateral trade reallocation effects. ### Key statistics and exact figures (preserved) - Dataset covers: 132 countries; years 1990 to 2023. - Contact period: June 2024 to July 2025; contacted around 120 countries. - IMF BPM seventh edition published: March 2025; recommends reporting invoicing currency as a supplementary item from 2029-30 onward. - Improvements versus Boz et al. (2022): about 20% more countries for dollar and euro invoicing patterns; renminbi invoicing patterns for more than 100 countries. - SAFE settlement data for China: renminbi settlement share of China’s exports reached 50% in 2023; US dollar share fell from 90% to 50%. - Renminbi invoicing magnitudes (2023 example): - 1.5% of total imports of countries in dataset invoiced in renminbi = USD 216 billion. - China exports invoiced in renminbi (if matched to settlement) = USD 1,646 billion. - Required invoicing in renminbi for countries not in dataset = USD 1,430 billion versus actual imports from China USD 1,373 billion → implied 104% of their imports from China; 18% of their total world imports. - Implied China export invoicing share under conservative assumptions = 6.5% vs SAFE settlement share 50%. - Maximum country coverage for US dollar export invoicing data in a year: 121 countries (in 2022). - After interpolation/backpolation/extrapolation dataset covers around 75% of global exports. - EU share of covered global exports: roughly one-third. ### Conclusion — synthesized findings - Dataset innovations: - Extension of trade invoicing currency shares to cover the years 2020–2023. - Incorporation of data on the Chinese renminbi. - Expanded coverage to 132 countries and revisions of earlier estimates. - Empirical synthesis: - Invoicing currency patterns have remained broadly stable globally in recent years, but important shifts exist. - Renminbi’s share in global trade invoicing remains modest but has grown rapidly since the early 2010s, initially concentrated in Asia and later expanding to many regions. - Geopolitical distance is an increasingly important correlate of invoicing currency choices, especially following Russia’s invasion of Ukraine in 2022: - Over full sample, dollar and euro use show little systematic relationship with geopolitical alignment; renminbi used more by countries closer to China. - Since 2022, use of the dollar and the euro are more negatively correlated with geopolitical distance from the US and the euro area. - Renminbi, home and third-country currencies increasingly supplant the dollar and euro in countries that have distanced themselves from the US and the euro area. - In countries moving closer to China, the dollar is being replaced by home and third-country currencies. - Overall: evidence points to an emerging fragmentation in invoicing patterns along geopolitical lines while confirming the resilience of a dominant currency. *Source: wpiea2025178-source-pdf - Section 6 concludes.* ### Section 6 concludes. ### Section 6 concludes. ### Updated and expanded dataset of trade invoicing currencies — scope and methodology - Dataset: unbalanced panel of invoicing currency shares in imports and exports of 132 countries from 1990 to 2023; invoicing currency shares generally refer to goods trade. - Country outreach and data collection: - Contacted national authorities of around 120 countries between June 2024 and July 2025. - Primary contacts: governors’ offices and senior officials in statistics, payments and international departments of central banks; subsequently statistics offices, ministries of finance, customs/revenue authorities when needed. - For more than half of the countries in the dataset, data were obtained by asking national authorities to compile information from raw customs records. - EU data sources (fourfold): - Annual ECB IRE data collection exercise (as in Boz et al. (2022)). - Non-public ECB archive data not reported in recent IRE exercises (as in Boz et al. (2022)). - Eurostat data (as in Boz et al. (2022)). - New: direct requests to national central banks (via the Eurosystem’s Working Group of External Statistics) to obtain invoicing currency information—especially for the renminbi (CNY). - Non-EU approach: systematic online search for official data; where not publicly available, direct requests to national authorities leveraging contacts across IMF, ECB, EBRD, AfDB, AACB, ADB, BIS, SEACEN. - Harmonization challenges: large cross-country heterogeneity in definitions, coverage (currencies, goods vs. services), reporting procedures, and whether customs forms record invoicing vs. settlement currency. ### Data properties, definitions, and treatment of missing observations - No global harmonized standard existed for reporting invoicing currency until the IMF BPM seventh edition published in March 2025; the IMF BPM encourages reporting invoicing currency as a supplementary item but only from 2029-30 onward. - In some cases, settlement currency is used as a proxy for invoicing currency following prior research (e.g., Gopinath 2015) when invoicing currency is unavailable. - To construct a balanced panel for aggregate trend analysis: - Linearly interpolate between missing observations. - Backpolate before the first available observation by assuming earlier observations had the same value as the earliest available. - Extrapolate after the last available observation by assuming later observations had the same value as the latest available. - Note: this simple extrapolation procedure stacks the deck against finding secular trends (e.g., rise in renminbi, decline in dollar invoicing). - After interpolation, backpolation and extrapolation, the dataset covers around 75% of global exports. - Maximum country coverage for dollar export invoicing in any year: 121 countries (in 2022). - EU countries account for roughly one-third of the share of global exports covered by the dataset. ### Country coverage limitations and notable omissions - Dataset does not include several important advanced and developing countries for various reasons: - Customs authorities do not record invoicing currency or do not record it with sufficient accuracy to allow publication; settlement currency proxy unavailable (examples: Hong Kong, Singapore, Mexico). - Unable to establish contact with relevant authorities (examples: Nigeria, Vietnam). - Only historical observations available from Boz et al. (2022) with no updates (examples: Ghana, Malawi, Tanzania). - Limited or single historical observations due to data storage changes or lack of access (examples: Canada — single observation for 2001 from Kamps (2006); India — no information after 2014 due to technical infrastructure changes; Pakistan — confined to 2001-2003 data from Kamps (2006)). - Geographic gaps: relatively limited information for Central America and Sub-Saharan Africa. ### Comparative coverage versus Boz et al. (2022) - Country and currency coverage improvement: - Information on US dollar and euro invoicing patterns for about 20% more countries compared to Boz et al. (2022). - Additional information for 2020-23 for generally all countries. - Renminbi invoicing patterns: information for more than 100 countries (substantially expanded relative to Boz et al. (2022)). - Note on figure interpretation: total dataset includes 132 countries but export/import country counts by currency may be lower because some countries report only imports or only exports for a currency. ### Is China’s settlement currency a useful proxy for invoicing currency? - Context: - PBoC measures to promote the renminbi in trade settlement: pilot program launched in 2009 (expanded in 2012), offshore renminbi markets in Hong Kong, London, Singapore; swap agreements; Shanghai Petroleum and Natural Gas Exchange launched in 2018; policy emphasis via Belt and Road Initiative. - SAFE settlement data (China): - Share of China’s exports settled in renminbi rose to 50% in 2023; US dollar share declined from 90% to 50% (SAFE data). - Empirical plausibility checks conducted using the updated dataset: - Back-of-the-envelope 2023 calculation: - About 1.5% of total imports of the countries in the dataset are invoiced in renminbi, amounting to about USD 216 billion. - If China’s trade settlement currency corresponded to trade invoicing currency, then USD 1,646 billion of China’s exports would be invoiced in renminbi. - This implies USD 1,430 billion of imports of countries not in the dataset would need to be invoiced in renminbi. - Comparison to actual: countries not in dataset import USD 1,373 billion from China, so the required USD 1,430 billion implies 104% of their total imports from China must be invoiced in renminbi. - As a share of their total imports from the world this would be 18%, versus 1.5% for total imports of countries in the dataset — a discrepancy judged implausible. - Implied share exercise (alternative assumption): if countries not in dataset invoice the same share of imports in renminbi as countries in dataset (and renminbi invoicing occurs only in bilateral trade with China), then: - Implied share of China’s exports invoiced in renminbi would be 6.5%, which is much lower than the 50% settlement share reported in SAFE data. - Analogous results for China’s imports produce similarly lower implied invoicing shares relative to SAFE settlement shares. - Conclusion on proxy quality: - Given the implausibly large adjustments required and the large gap between implied invoicing shares and SAFE settlement shares, settlement currency (SAFE) is not used as a proxy for invoicing currency in China’s trade in this dataset. - Consistent with Boz et al. (2022), SAFE trade settlement currency data are not included as a proxy for trade invoicing currency for China. ### Key statistics and exact figures highlighted - Dataset covers: 132 countries; years 1990 to 2023. - Contact period: June 2024 to July 2025; contacted around 120 countries. - IMF BPM seventh edition published: March 2025; recommends reporting invoicing currency as a supplementary item from 2029-30 onward. - Improvements versus Boz et al. (2022): about 20% more countries for dollar and euro invoicing patterns; renminbi invoicing patterns for more than 100 countries. - SAFE settlement data for China: renminbi settlement share of China’s exports reached 50% in 2023; US dollar share fell from 90% to 50%. - Renminbi invoicing magnitudes (2023 example): - 1.5% of total imports of countries in dataset invoiced in renminbi = USD 216 billion. - China exports invoiced in renminbi (if matched to settlement) = USD 1,646 billion. - Required invoicing in renminbi for countries not in dataset = USD 1,430 billion versus actual imports from China USD 1,373 billion → implied 104% of their imports from China; 18% of their total world imports. - Implied China export invoicing share under conservative assumptions = 6.5% vs SAFE settlement share 50%. - Maximum country coverage for US dollar export invoicing data in a year: 121 countries (in 2022). - After interpolation/backpolation/extrapolation dataset covers around 75% of global exports. - EU share of covered global exports: roughly one-third. *Source: wpiea2025178-source-pdf - Section 6 concludes.* ### 3.1 The US dollar and the euro ### 3.1 The US dollar and the euro ### Evolution of export and invoicing shares - The combined share of dollar and euro invoicing has remained fairly stable over time. - The dollar is the dominant vehicle currency in global trade: the share of exports invoiced in dollars far exceeds the share of exports to the US. - The share of global exports invoiced in euros is about as large as the share of global exports to the euro area, consistent with the euro being used mainly in trade with the euro area rather than as a vehicle currency in third-country trade. - There are no significant shifts in dollar and euro invoicing shares at the global level from 2020 to 2023—the four additional years in the dataset compared to Boz et al. (2022). ### Robustness when excluding issuer-area countries - When euro area countries (and the US) are excluded: - The share of the dollar in global trade becomes an even greater multiple of the share of global exports destined to the US. - The share of the euro in trade invoicing is almost halved and is—at best—about as large as non-euro area countries’ exports to the euro area. - Dropping euro area countries reveals global euro invoicing and exports to the euro area have been converging, indicating an increasing importance of the euro over time; this growth in euro invoicing can be traced to non-euro area European countries. ### Data treatment note on pegged currencies - Invoicing in currencies pegged to the dollar or the euro is accounted for by adding invoicing in home currency to invoicing in dollar/euro for countries whose currencies are pegged to the dollar/euro. - Peg classification: a currency is classified as pegged in a given year if the fine assessment of Ilzetzki et al. (2019) is “1: No separate legal tender or currency union”, “2: Pre-announced peg or currency board arrangement”, “3: Pre-announced horizontal band that is narrower than or equal to +/-2%”, or “4: De facto peg”. A currency is classified as having been pegged over the full sample period if it was pegged for every year since 2005 until 2023. - Countries for which home currency is added to dollar invoicing shares: Bahrain, Brunei Darussalam, Curacao, Jordan, Macao, Maldives, Saudi Arabia and Sint Maarten; except for Saudi Arabia, the home currency shares are low. - Countries for which home currency is added to euro invoicing shares: Bosnia and Herzegovina, and the Republic of Congo. ### 3.2 Enters the renminbi ### Renminbi invoicing trends - The share of global exports to China has increased strongly since 2000 and is now comparable to the share of exports to the US. - The share of global exports invoiced in renminbi is hardly visible when assessed on the same scale as the dollar and the euro. - Using raw (rather than inter/extrapolated) renminbi invoicing shares to capture recent strong growth: - China has become an increasingly important export market for countries in all regions, especially in Asia, Africa, and Latin America. - Renminbi invoicing was hardly visible until 2010; since then it has increased across countries in most regions, particularly in Asia and more recently in Europe and Latin America. - Despite recent strong increases, renminbi invoicing remains below the share of global exports to China. - Figure A.7 documents that zoomed conclusions are very similar when looking at unweighted averages across countries and when excluding euro area countries or Russia. ### 4 Invoicing currency patterns and geopolitics ### Theoretical channels linking geopolitics to invoicing - Standard models emphasize export patterns, input-output linkages, and competition with foreign and local producers as determinants of exporters’ choice of invoicing currency; exporters choose the currency that makes optimal price deviations least volatile considering future shocks to demand and input costs. - Geopolitical alignment may influence invoicing currency patterns through: - Export-reorientation channel: tariffs, subsidies, moral suasion, or trade sanctions can redirect exports and hence change invoicing currency (e.g., from dollar-invoiced exports to the US toward euro-invoiced exports to the euro area). - Marginal cost / input substitution channel: trade barriers or "friend-shoring" can prompt substitution of intermediate inputs (e.g., replacing US dollar–invoiced inputs with euro-invoiced inputs), reducing desired-price volatility and strengthening incentives to invoice in the new input currency. - Complementarity and spillover channel: shifts in safe asset currency holdings and foreign exchange reserve rebalancing towards geopolitically aligned currencies may favor analogous switches in trade invoicing currency; increased costs for cross-border payments (e.g., punitive measures associated with sanctions) can reduce dollar trade invoicing in affected trade relationships. ### Empirical approach preview - The dataset is used to explore relationships between geopolitics and trade invoicing currency patterns at the global level. - Geopolitical alignment is measured using the ideal point distance metric based on UN General Assembly voting patterns (Bailey et al. 2017). - Classification: the upper (lower) 25% percentile in the distribution of the distance to the US in 2023 across countries is defined as US most-aligned (US least-aligned); a US neutral group accounts for 50% of countries. ### 4.2 Invoicing currencies and geopolitical alignment ### Cross-country export shares by alignment (sample shares) - With this classification, in the country sample: - 34% of world exports originate from US most-aligned countries. - 8% of world exports originate from US least-aligned countries. - 9% of world exports originate from the US itself. - 49% of world exports originate from US neutral countries. ### Key empirical findings on the dollar - When comparing country-group shares: - A significantly larger share of total exports is invoiced in dollars in US least-aligned countries compared to US most-aligned countries, despite both groups having similar shares of exports to the US. This suggests the dollar plays a more prominent role as a vehicle currency in US least-aligned countries. - US least-aligned countries are predominantly emerging market and developing economies, which tend to invoice in dollars more frequently than advanced economies. - Using an invoicing currency may be more readily abandoned in third-country trade than in bilateral trade with the currency issuer, implying greater potential for a shift away from the dollar toward a challenger currency in US least-aligned countries; structural factors may limit the practical ability to reduce dollar reliance. - Dollar-invoiced trade has declined since the early 2010s in US least-aligned countries, while it has been stable in US most-aligned countries. - Exports from least-aligned countries to the US have experienced a secular decline, but the decline in dollar invoicing has been larger—supporting that the decline primarily reflects a reduction in vehicle-currency–invoiced trade. ### Key empirical findings on the euro - Contrasting with the dollar: - The share of exports invoiced in euros is substantially lower in euro area least-aligned countries than in euro area most-aligned countries. - The euro plays no role as an invoicing—let alone as vehicle—currency in euro area least-aligned countries, based on the comparison of invoiced shares and exports to the euro area. - The share of exports invoiced in euros in euro area most-aligned countries has been increasing; until around 2010 that increase was even faster than the corresponding share of exports, consistent with increasing euro use in euro area and EU neighboring countries. *Source: wpiea2025178-source-pdf - 3.1 The US dollar and the euro* ### 6. To do so, we plot the share of exports invoiced in dollars over time for individual US least-aligned ### 6. To do so, we plot the share of exports invoiced in dollars over time for individual US least-aligned countries ### Decomposition of dollar export invoicing among US least-aligned countries - Figure 7 decomposes the evolution of the share of country-group exports invoiced in dollars by US least-aligned countries: - Left panel: stacked contributions by country to the overall share of exports invoiced in dollars (group-level series shown in top left panel of Figure 6). - Right panel: unstacked country-level shares. - Primary drivers of the group-level decline in dollar invoicing (green dashed line in top left panel of Figure 6) are Russia and Saudi Arabia. - Note: Figure A.8 presents the analogous decomposition for the top-right panel of Figure 6. ### Renminbi invoicing and trade with China by geopolitical alignment - Figure 8 shows evolution of: - Share of country-group imports from China (left panel). - Share of country-group imports invoiced in renminbi (right panel). - Data choices and caveats: - Imports are used rather than exports because renminbi invoicing currency data are more numerous and longer for imports (especially for Russia). - The series use raw renminbi invoicing shares rather than inter/extrapolated data. - Vertical lines in the figures indicate 2022. - Empirical patterns: - China has been gaining importance as an import source, especially for US least-aligned countries, with divergence after the recovery from the disproportionate drop in trade with China during the COVID-19 pandemic in 2020 and Russia’s invasion of Ukraine in 2022. - Renminbi invoicing has "taken off" in US least-aligned countries; the increase in US most-aligned countries has been modest and has even reversed since Russia’s invasion of Ukraine in 2022. - The divergence in both import shares to China and renminbi invoicing shares is particularly stark when Russia is included in the US least-aligned group (green dash-dotted lines, right-hand side axis), but remains visible when Russia is dropped (green dashed lines, left-hand side axis). - Figure A.10 decomposes renminbi invoicing growth into country contributions, showing that among US most-aligned countries the increase was mostly due to European countries, while among US least-aligned countries it was mostly due to Asian countries. ### Key takeaways (five) - First, US least-aligned countries tend to invoice a larger share of their exports in dollars compared to US most-aligned countries. - Second, because US least-aligned countries account for a smaller share of global trade, the value of their exports invoiced in dollars remains lower than that of US most-aligned countries. - Third, while the share of exports by US least-aligned countries invoiced in dollars has been declining since the early 2010s, this long-term trend has not been broad-based and has been driven primarily by Russia and Saudi Arabia. - Fourth, in contrast to the dollar, euro invoicing in global exports is predominantly by euro area most-aligned countries. - Fifth, trade with China and renminbi invoicing have been growing faster for US least-aligned countries since 2014, and at least since 2021 not only due to Russia. ### Changes in geopolitical distance and invoicing currency patterns (country-level evidence) - Approach: - Use continuous geopolitical distance measure of Bailey et al. (2017) (ideal point distance and UN General Assembly voting patterns) rather than categorical most/least-aligned groups. - Compare changes between 2015-19 and 2022-23 (post-invasion period). - Figure 9 (changes in bilateral geopolitical distances between 2015-19 and 2022-23): - Left panel: changes in geopolitical distance to the US (vertical axis) versus to China (horizontal axis); euro area countries are dropped in left panel. - Right panel: changes in geopolitical distance to the euro area (vertical axis) versus to China (horizontal axis); geopolitical distance to the euro area measured as unweighted average of distances to individual euro area countries. - Observed shifts: - Majority of countries moved closer geopolitically to the US and away from China (bottom-right quadrant of left panel). - A small group—including Russia and Belarus—moved farther from the US and closer to China (top-left quadrant). - Among countries that moved closer to the US and further from China, most also became more geopolitically distant from the euro area. - Only a few countries—Russia, Belarus, Kazakhstan, Uzbekistan and Mali—distanced themselves from both the US and the euro area. - Dollar invoicing changes versus geopolitical distancing: - Figure 10: - Left panel: countries with the largest drops in dollar export invoicing currency shares in 2022-23 relative to the 2015-19 average include Kyrgyzstan, Russia, Belarus, and Uzbekistan. - Data note: for Russia, Belarus, Kazakhstan and Kyrgyzstan, trade settlement currency is used rather than invoicing currency shares due to data limitations (see footnote 16). - Right panel: negative association between increases in geopolitical distance from the US and declines in dollar export invoicing shares across the full country sample. - Russia, Belarus, Kyrgyzstan, and Uzbekistan exhibit particularly large shifts in both geopolitical distance and dollar invoicing. - Caveat: Figure 10 does not by itself distinguish whether declines in dollar invoicing reflect geopolitical considerations beyond bilateral trade composition effects. ### Illustrative country case — Russia - Figure 11 (imports and invoicing shares for Russia): - Left panel: share of Russia’s imports sourced from China (blue solid line, left axis) and share of Russia’s imports invoiced in renminbi (red dashed line, right axis). - Right panel: share of Russia’s imports sourced from the euro area (blue solid line, left axis) and share invoiced in euro (red dashed line, right axis). - Interpretation: - Changes in renminbi and euro invoicing shares closely track shifts in bilateral trade shares. - The rise in renminbi invoicing in Russia primarily reflects a reorientation of imports being sourced from China and those imports being invoiced in renminbi. - The decline in euro invoicing corresponds to reduced euro-invoiced trade with the euro area. - Data and proxy issues: - For Russia we use trade settlement currency as a proxy for invoicing currency. - Comparison between Russia’s trade settlement currency data from the Central Bank of Russia and trade invoicing currency data from Russia’s customs (as used in Chupilkin et al. (2023) and shown in Figure A.11) suggests the two series are very similar or follow comparable trends over time for the dollar, euro and renminbi. - There is evidence that the Central Bank of Russia drew from its swap line with the PBoC to provide renminbi liquidity to domestic banks so they could pay for imports from China and preserve dollar reserves (see Horn et al. 2023, Chupilkin et al. 2023), which may affect settlement versus invoicing distinctions. ### Next analytical steps signaled in the text - The authors indicate they will use regression analysis to further explore the role of developments in commodity trade and geopolitical alignment for the decline in dollar invoicing (referenced after Figure 7). - They also propose to examine whether invoicing currency shifts have broadened beyond Russia and Saudi Arabia since Russia’s invasion of Ukraine in 2022, and whether changes go beyond compositional effects linked to bilateral trade reallocation. *Source: wpiea2025178-source-pdf (chapter/section content).* ### 4.4 Evidence from panel regressions ### 4.4 Evidence from panel regressions ### Methodology - Sample period: 1999-2023. - Main specification (Equation (1)) estimates exporter-level export invoicing currency shares s^c_{e,t} using: - lagged dependent variable s^c_{e,t-1}; - vector of controls w^c_{e,t} including: share of exporter’s exports to the issuer of currency c (US, euro area, China), share of exports to economies that peg to that issuer (for renminbi use share of trade with ASEAN countries), commodity/export-in-oil shares, and bilateral nominal exchange rate between exporter’s currency and the dollar/euro/renminbi; - exporter and time fixed effects α^c_e and τ^c_t. - Regularization: exclude observations when first difference over time falls below the 1st percentile or exceeds the 99th percentile. - Inference: Driscoll-Kraay standard errors robust to serial correlation and cross-section dependence. - Dependent variable: share of total exports invoiced in each currency (bilateral invoicing data not available). - Alternative and pooled specifications (Equations (2)–(4)) assess geopolitical distance effects and interactions with a post-2021 dummy (I_{t>2021}). ### Summary statistics (selected, as reported in Table 1) - USD invoicing share: mean 49.15; min 0.45; p5 5.96; p50 42.96; p95 98.61; max 100.00; sd 33.33; count 1,307 - EUR invoicing share: mean 41.24; min 0.00; p5 0.29; p50 39.44; p95 91.94; max 99.46; sd 35.26; count 1,303 - CNY invoicing share: mean 0.25; min 0.00; p5 0.00; p50 0.00; p95 1.17; max 29.32; sd 1.40; count 704 - Share of trade with US: mean 7.63; min 0.00; p5 0.45; p50 4.59; p95 27.73; max 44.73; sd 8.41; count 1,307 - Share of trade with euro area: mean 34.65; min 0.05; p5 3.04; p50 38.43; p95 67.59; max 80.68; sd 21.71; count 1,303 - Share of trade with China: mean 7.24; min 0.00; p5 0.21; p50 2.57; p95 26.52; max 92.60; sd 10.36; count 704 - Commodity trade share: mean 23.40; min 0.02; p5 2.87; p50 14.04; p95 74.80; max 98.58; sd 22.69; count 1,250 - Oil trade share: mean 6.38; min 0.00; p5 0.00; p50 2.22; p95 34.30; max 55.58; sd 10.18; count 1,307 - Bilateral exchange USD exchange rate (log): mean 440.52; min 111.37; p5 366.41; p50 449.91; p95 476.33; max 498.24; sd 39.29; count 1,307 ### Main regression findings — compositional and mechanical correlates (Table 2) - Persistence: Lagged invoicing currency share coefficients (columns reported for USD, EUR, CNY) are large and precisely estimated: 0.73, 0.75, 0.80, 0.80, 0.92, 1.02 (all with (0.00) shown for associated p-values). - Bilateral trade share with invoicing currency issuer: - Coefficients: 0.11, 0.06, 0.08, 0.09, 0.01, 0.01 - Parenthetical p-values present: (0.00)(0.07)(0.00)(0.00)(0.00)(0.00) - Interpretation: share of exports destined for the issuer is positively correlated with share invoiced in that issuer’s currency; effect largest for the dollar, smaller for the euro, smallest for the renminbi. - Commodity and oil trade: - Commodity trade share coefficients: 0.09, -0.04, -0.00 with p-values (0.00)(0.04)(0.74) - Oil trade share coefficients: 0.17, -0.16, -0.01 with p-values (0.01)(0.00)(0.04) - Interpretation: commodity/oil trade is positively correlated with dollar invoicing and negatively correlated with euro and renminbi invoicing, consistent with commodities/oil being predominantly invoiced in dollars. - Bilateral exchange rate controls show small and mixed coefficients. - Within R-squared across columns: 0.65, 0.65, 0.77, 0.78, 0.80, 0.96 (observations and country counts reported per column in Table 2). ### Geopolitical distance effects — average and post-2021 changes (Table 3) - Baseline (full sample) correlations: - On average over the entire sample, dollar and euro invoicing use are not systematically related to geopolitical distance to their issuers. - Renminbi: use tends to be greater for exporters geopolitically closer to China (consistent with Figure 8). - Interaction with post-2021 dummy (Russia’s invasion of Ukraine, 2022): - Geopolitical distance to invoicing currency issuer coefficients (columns (1)–(6)): 1.06, -0.51, 0.09, -0.28, -0.08, -0.05 (parenthetical p-values for these estimates shown in table). - Geopolitical distance to invoicing currency issuer × post-2021 coefficients: -0.12, -0.36, -0.44, -0.77, -0.05, -0.03 with associated parenthetical p-values (0.68)(0.09)(0.06)(0.00)(0.14)(0.11). - p-value full effect (overall effect including interaction) reported per column: 0.13, 0.34, 0.52, 0.08, 0.01, 0.04. - Interpretation: - Since 2022, the correlation between dollar and euro invoicing and geopolitical distance to their issuers has become more negative. - For the euro the overall negative effect since 2022 is statistically significant at conventional levels (see p-value full effect). - For the renminbi the negative correlation with geopolitical distance from China has become more negative since 2022; significance depends on commodity vs oil controls (column (6) shows statistical significance). ### Geopolitics and invoicing across issuers and currencies (Table 4; condensed) - Specification: Equation (4) estimates effects for all combinations of invoicing currency c (USD, EUR, CNY, Home, Home (incl. EUR), Other) and geopolitical distance to issuers i ∈ {US, EA, CHN}, controlling separately for commodity share and oil share. - Key reported coefficients and patterns (selected numbers reported exactly as in the source text): - Geopolitical distance to the US × post-2021, oil-share control, column (8): dollar −0.36 (parenthetical p-value shown as (0.09) in table context) and for geopolitical distance to China × post-2021, column (8): 0.86 (parenthetical p-value (0.00) reported). - Summary pattern: since 2021/2022 - Greater geopolitical distance from the US is associated with a lower use of the dollar as an export invoicing currency (negative post-2021 interaction estimates). - Geopolitical distancing from the US is associated with greater use of the renminbi, home, and other third currencies (green cells in Figure 12 summary). - Geopolitical distancing from the euro area reduces the use of the euro in export invoicing and is associated with increases in renminbi, home, and other currencies. - Geopolitical distancing from China (when controlling for oil share) is associated with a greater tendency to invoice in dollars, while use of home or other third-party currencies declines; when controlling for commodity share, dollar and renminbi use seem not to change significantly with distancing from China based on conventional significance levels. - Intuition: countries moving geopolitically closer to the US have tended also to draw nearer to the euro area and distance from China; invoicing currency choices have rotated along these geopolitical lines since 2022. ### Robustness and additional checks - Results are similar when using import invoicing currency shares as dependent variable (Table B.1), and when using trading-partner average geopolitical distance (Table B.3). - Figure A.15 shows time variation in the conditional correlation, confirming it has become (more) negative recently. - Table B.4 documents that results for euro and renminbi export invoicing in Table 3 are not driven by Russia, Belarus or Uzbekistan; for dollar export invoicing shares coefficients of geopolitical distance to the US can turn insignificant when excluding these countries (data availability limits noted). - Table B.5: results for euro invoicing shares robust to alternative definitions of geopolitical distance to the euro area. - Pooled exporter–invoicing currency panel regressions (Equation (2)) reported in the appendix impose homogeneity across currencies and can yield different coefficient estimates; the authors focus on currency-specific regressions due to heterogeneous effects. ### Key takeaways - Compositional factors matter: bilateral trade with the invoicing currency issuer and commodity/oil trade are important correlates of invoicing currency shares; commodity/oil trade strongly supports dollar invoicing. - Persistence is strong: lagged invoicing shares explain a large fraction of variation. - Geopolitics matter increasingly since 2022: - The correlation between the use of the dollar and the euro for invoicing and geopolitical distance to their issuers has become more negative since Russia’s invasion of Ukraine. - Renminbi use tends to be higher for exporters closer to China and has become more negatively correlated with distance from China since 2022 (statistical significance varies by specification). - Post-2021 patterns indicate a geopolitical rotation in invoicing: distancing from the US/euro area is associated with increased use of renminbi, home, and other currencies; distancing from China (in some specifications) is associated with increased dollar invoicing. *Source: wpiea2025178-source-pdf* ### 5.1 Invoicing currency choice for commodity trade ### 5.1 Invoicing currency choice for commodity trade ### Key mechanisms and theoretical considerations - Commodity products—particularly oil—have been predominantly invoiced in dollars in recent decades (Eichengreen et al. 2016, McLeay & Tenreyro forthcoming). - Homogeneous nature of commodity products and limited pricing power of producers makes it convenient to quote prices in a single currency for ease of comparison. - Commodities are typically traded on organized exchanges; prices adjust rapidly to global supply and demand conditions, with arbitrage quickly eliminating deviations from world prices. - The invoicing currency also serves as the settlement currency, raising interactions between invoicing and settlement currency choices, which can be important in practice for trade between economies aligned with different geopolitical blocs. - Theoretical insights: - In standard models, choice of invoicing currency is typically inconsequential when prices are flexible. - The invoicing currency of intermediate inputs can influence an exporter’s invoicing decision through the marginal costs channel (Gopinath et al. 2010, Mukhin 2022). - Recent work shows exchange rates can have allocative effects even when commodity dollar prices are flexible (McLeay & Tenreyro forthcoming): if local input factor prices are sticky in domestic currency, firms in a small open economy exporting commodities experience a rise in profits expressed in dollar terms and raise production following exchange rate depreciation vis-`a-vis the dollar. - Settlement currency restrictions (e.g., prohibitions, sanctions) can reduce the attractiveness of a currency for invoicing by increasing exchange rate risk and deviations from the optimal preset price. ### 5.2 Reduced use of the dollar in oil export invoicing of US least-aligned countries? ### Anecdotal evidence of alternative settlement currencies - Reported cases of non-dollar settlement: - Russia has settled oil exports to China in renminbi (Reuters 2023c). - Iran and Venezuela believed to have shifted to renminbi for settling oil exports to China (CNN 2012, Reuters 2019). - Venezuela reported to increasingly settle oil exports in cryptocurrencies (Reuters 2024). - India has settled oil imports from the United Arab Emirates in Indian rupees (Reuters 2023a). - India settled some oil imports from Russia in renminbi and has discussed settling in rubles (Reuters 2023b; Bloomberg 2024). - Reports of negotiations to settle a portion of Saudi Arabia’s oil exports to China in renminbi (Wall Street Journal 2022). - Many of these countries—such as Russia, Venezuela, Iran, and China—are not geopolitically aligned with the US. ### Data limitations and analytic approach - Lack of currency-use data specifically for commodity transactions limits definitive conclusions. - The analysis is suggestive: examines whether evidence is symptomatic of reduced reliance on the dollar in commodity trade for US least-aligned exporters. - Focus is on oil (rather than broader commodities) because oil is more homogeneous and anecdotal evidence pertains primarily to oil. ### Descriptive patterns: world oil prices and oil trade shares - Historically, oil prices have co-moved strongly with oil trade shares—largely mechanical when invoiced in dollars: when oil price rises, the value of oil trade quoted in local currency increases, raising its share in total trade (absent short-run demand or exchange rate adjustments). - A weakening of this co-movement may indicate a shift away from dollar invoicing, with exchange rate fluctuations contributing to decoupling. - Figure 14 (described): - For Russia: recent weakening of co-movement between world oil prices and Russia’s oil export share—consistent with a shift away from dollar invoicing, though an alternative explanation is the G7 price cap detaching Russian pricing from global benchmarks. - For Canada (benchmark oil producer geopolitically aligned with the US): oil export share continues to exhibit strong co-movement with world oil prices, suggesting Russia’s decoupling may be idiosyncratic. ### Regression analysis: specification and interpretation - Regression estimated (as presented): s_usd_{e,t} = ρ s_usd_{e,t-1} + β_1 w_usd_{e,t} + γ r ω_{oilx}_{e,t} ˆI_{pt>2021q} ˆI_{p∈RA_US q} + δ_1 z_{e,t} + α_e + τ_t + u_{e,t}, where z_{e,t} includes all other interaction terms and A_US is the set of US most-aligned countries. - Bilateral exchange rates against the US dollar and time fixed effects included to control for alternative explanations (e.g., oil price declines relative to non-oil export prices for US least-aligned countries, oil demand falls relative to non-oil demand, currencies of US least-aligned countries depreciate less against the dollar compared to those of US most-aligned countries). ### Regression results (Table 5) - Coefficient of interest γ captures whether correlation between commodity exports and dollar invoicing shares declined for US least-aligned economies after 2021. - Main estimates (columns (1)–(3)): - Oil trade share coefficients: - Column (1): 0.18 (standard error reported as (0.00)) - Column (2): 0.25 (standard error reported as (0.00)) - Column (3): 0.26 (standard error reported as (0.00)) - Oil trade shareˆleast-alignedˆpost-2021 interaction coefficients: - Column (1): -0.04 (standard error (0.30)) - Column (2): -0.02 (standard error (0.49)) - Column (3): 0.03 (standard error (0.45)) - Trade to GDP (in columns (2)–(3)): -0.01 (standard errors (0.71), (0.46)) - Controls and specification notes: - Standard controls: ✓ in all columns - Full set of interactions: ✓ in all columns - Year FEs: ✓ in all columns - YearˆLeast-aligned FEs: included in column (3) only - Within R-squared: - Column (1): 0.65 - Column (2): 0.66 - Column (3): 0.65 - Observations: - Column (1): 1307 - Column (2): 1283 - Column (3): 1262 - Countries: - Column (1): 111 - Column (2): 107 - Column (3): 105 - Note: Column (3) excludes Belarus and Russia relative to column (2). - Interpretation: - The estimated γ is negative in column (1), consistent with a declining correlation between oil export shares and dollar invoicing shares, but it is estimated imprecisely. - Results remain similar when controlling for overall exports and including separate year fixed effects for US least-aligned countries, and when excluding Belarus and Russia. - Conclusion: impossible to ascertain whether negative sign reflects a true decline in correlation or noise in the data. - Robustness: - Results based on commodity instead of oil export share (Table B.8) are consistent with Table 5, with somewhat more precise coefficients, but no anecdotal evidence of comparable policy efforts targeting de-dollarization in non-oil commodities. ### 6 Conclusion (relevant excerpt) - The paper updates and extends the dataset on global trade invoicing currency patterns originally compiled by Boz et al. (2022). - Key dataset innovations: - Extension of trade invoicing currency shares to cover the years 2020–2023. - Incorporation of data on the Chinese renminbi. - Expanded coverage to 132 countries and revisions of several earlier estimates. - Main empirical findings summarized: - Invoicing currency patterns have remained broadly stable globally in recent years, but important shifts exist. - The renminbi’s share in global trade invoicing remains modest but has grown rapidly since the early 2010s, initially concentrated in Asia and later expanding to many regions. - Geopolitical distance is an increasingly important correlate of invoicing currency choices, especially following Russia’s invasion of Ukraine in 2022: - Over the full sample, use of the dollar and the euro exhibits little systematic relationship with geopolitical alignment, while the renminbi is generally used more by countries geopolitically closer to China. - Since 2022, the use of the dollar and the euro are more negatively correlated with geopolitical distance from the US and the euro area. - The renminbi—along with home and third-country currencies—increasingly supplants the dollar and the euro in countries that have distanced themselves geopolitically from the US and the euro area. - In countries that have moved closer to China, the dollar is being replaced by home and third-country currencies. - These findings point to an emerging fragmentation in invoicing patterns along geopolitical lines, while confirming the resilience of a dominant currency. *Source: IMF Working Paper chapter section 5.1–5.2 and concluding excerpts from the provided PDF content.* ### Introduction of the Euro as a Natural Experiment’,Journal of International Economics150(C). ### Patterns of invoicing currency in global trade in a fragmented world economy ### Bibliography and related research cited - Extensive list of working papers, journal articles, policy notes, and media reports cited, including but not limited to: - Studies on currency invoicing, dollar dominance, renminbi internationalization, and geopolitical fragmentation (examples: Boz et al. (2022), Gopinath et al. (2020), Eichengreen (2011), Chinn & Frankel (2008), Clayton et al. (2024, 2025)). - Empirical pieces and working papers on sanctions, trade invoicing, settlement, and geoeconomic fragmentation (examples: Berthou (2023), Chupilkin et al. (2023), Fernandez-Villaverde et al. (2024)). - Central bank and media sources on renminbi/other currency settlements (examples: People’s Bank of China (2012), Centralbanking.com (2023), Reuters, Bloomberg, CNN). - Methodological and data contributions cited include Papke & Wooldridge (1996) for econometric methods for fractional response variables, and SAFE (data source referenced in figure notes). ### Evolution of invoicing and settlement currencies (figures and captions) - China-specific settlement and implied invoicing patterns: - Figure A.1: Evolution of the share of China’s trade settled in renminbi (CNY, red dashed), US dollar (USD, green solid), and euro (EUR, blue dash-dotted) for exports and imports over 2010–2020; data obtained from SAFE. - Figure A.2: Implied renminbi import and export invoicing shares for countries not in the dataset derived from SAFE CNY export settlement shares under assumptions that (i) China’s exports settled in renminbi are also invoiced in renminbi and (ii) renminbi invoicing occurs only in bilateral trade with China. (Panels cover 2010–2020 and reference Appendix C for calculation details.) - Coverage and aggregation of invoicing data: - Figure A.3: Country and world export share coverage for invoicing currency data plotted over 1990–2020 with left panels showing number of countries (lhs) and share of world exports (rhs); right panels show share of world exports covered after interpolation and extrapolation. Separate panels for US dollar, euro, and renminbi. - Global shares of trade vs. invoicing currency (aggregate patterns): - Figure A.4 (left): Shares of exports to the US, the EA, and RoW and shares of exports invoiced in USD, EUR, and other currencies (interpolated/extrapolated data averaged 1999–2023). - Figure A.4 (right): Same decomposition splitting USD-invoiced exports into commodity and non-commodity, assuming all commodity exports are invoiced in dollars and using World Bank WDI shares for commodity trade (agricultural raw materials, ores and metals, fuels). - Country-level trade and invoicing correlations: - Figure A.5: Scatter plots (averages over 1990–2023) of each country’s share of total exports accounted for by the US vs. share of exports invoiced in USD (left panel), and share accounted for by the euro area vs. share invoiced in EUR (right panel). Averages computed only for years with both invoicing and export share data. - Global time series excluding euro area countries: - Figure A.6: Evolution of share of exports to US, EA, China, and RoW in total global exports and corresponding shares of global exports invoiced in USD, EUR, CNY, and other currencies for 2000–2020, excluding euro area countries. Based on interpolated and extrapolated data. - Regional and specification robustness for renminbi invoicing: - Figure A.7: Multiple specifications of CNY invoicing shares over 2000–2020: - Top row: intra/extrapolated renminbi invoicing shares. - Second row: unweighted renminbi export invoicing shares (left) and export-share weighted invoicing shares (right). - Third row: same export-weighted shares excluding euro area countries. - Fourth row: excluding Russia from the European bloc. - Regional series shown for Asia, Europe, North America, Latin America, Africa, Middle East. - Decomposition of USD invoicing by non-aligned countries: - Figure A.8: Decomposition (by country) of evolution in the share of global exports invoiced in USD by US non-aligned countries over 2000–2020. Panels show stacked country contributions and series for selected country codes (examples appearing in legend: MYS, IDN, DZA, SAU, RUS). - China trade and CNY invoicing by geopolitical alignment: - Figure A.9: - Left panel: Share of exports destined to China for country-groups (US most-aligned, US least-aligned excluding RUS, US least-aligned including RUS) over 2014–2022. - Right panel: Renminbi invoicing shares for those country-groups over 2014–2022. Vertical lines mark 2022. - Decomposition of change in imports from China invoiced in renminbi: - Figure A.10: - Panels for US least-aligned countries and US most-aligned countries show contributions of top five countries within each group and aggregates of remaining countries to the share of imports from China invoiced in CNY over 2014–2022. Country codes illustrated include IDN, MYS, MNG, RUS, UZB, DEU, FRA, ITA, NLD, ROU, among others. - Raw renminbi invoicing shares used (not inter/extrapolated). - Settlement vs. invoicing currency comparison for Russia: - Figure A.11: - Series for 2016–2023 comparing settlement currency shares in Russia’s imports from the rest of the world (Central Bank of Russia data) with invoicing currency shares in Russia’s imports from the rest of the world excluding EEA (Chupilkin et al. (2023) customs-based invoicing data). - Panels show USD, EUR, CNY, RUB settlement/invoicing percentage series with specific panels depicting imports from RoW and imports from non-CIS countries; caption notes lines: blue solid (settlement, imports from RoW), red dash-dotted (settlement, imports from non-CIS), green dashed (invoicing, imports from RoW excluding Eurasian Economic Union). - Geopolitical distance distributions and networks: - Figure A.12: Density plots of cross-country distribution of geopolitical distance to the US (horizontal axis) and China (vertical axis) in 2013 and 2023 using Bailey et al. (2017) measures; brighter areas indicate greater mass of countries. - Figure A.13: Country network charts for bilateral geopolitical distances to the US and China for 2013 and 2023 (nodes labeled: European Union, United States, China, Other). ### Data and methodological notes (from figure captions) - SAFE used as the data source for China settlement shares. - Interpolation and extrapolation applied to invoicing currency coverage to produce world export coverage and time series (figures specify interpolated/extrapolated data and averaging windows, e.g., 1999–2023). - Assumptions for imputations: e.g., renminbi invoicing implied for missing countries under assumptions stated in Figure A.2 caption. - Commodity invoicing assumption: all commodity exports invoiced in dollars used in Figure A.4 decomposition, with commodity classification from World Bank “World Development” Indicators components: agricultural raw materials, ores and metals, and fuels. - Country-group alignments and country codes are used consistently in decompositions and panels (examples: US most-aligned, US least-aligned; country codes such as IDN, MYS, RUS, DEU, FRA, ITA, NLD, ROU, MNG, UZB). *Source: wpiea2025178-source-pdf - Introduction of the Euro as a Natural Experiment’,Journal of International Economics150(C).; Figures and captions from the Online Appendix to "Patterns of invoicing currency in global trade in a fragmented world economy" (authors: Emine Boz, Anja Brüggen, Camila Casas, Georgios Georgiadis, Gita Gopinath, Arnaud Mehl).* ### 2023. Geopolitical distance is given by the ideal distance point estimates from Bailey et al. (2017). Each node repres ### Patterns of Invoicing Currency in Global Trade in a Fragmenting World Economy — Working Paper No. WP/2025/178 (excerpts) ### Changes in bilateral geopolitical distance (Figures A.14 and A.16) - Figure A.14: - Shows changes in bilateral geopolitical distances to the US and the euro area between 2015-19 and 2022-23 for countries in the dataset. - Geopolitical distance is measured based on the ideal point distance of Bailey et al. (2017) and voting patterns in the UN General Assembly. - Horizontal axis: changes in geopolitical distance to the euro area. - Vertical axis: changes in geopolitical distance to the US. - Geopolitical distance to the euro area is measured as the unweighted average of distances to individual euro area countries. - Euro area countries are dropped from the figure. - Axis ticks and labels preserved: vertical axis range includes −1, −.5, 0, .5 with label "∆ Geopolitical distance to US"; horizontal axis label shows −.5 0 .5 1 with "∆ Geopolitical distance to EA". - Country labels plotted in the figure include (selected examples exactly as shown): CHL, IDN, MYS, UZB, THA, NAM, KHM, NZL, JPN, KGZ, KOR, ZWE, HUN, PHL, BGD, AUS, ISR, GEO, HRV, SRB, TUR, NOR, KEN, AGO, CZE, PNG, URY, BGR, ARM, ZMB, JOR, EGY, MDV, UKR, ARG, BWA, CRI, ALB, PRY, ISL, LBR, MKD, MDA, SYC, BLZ, KAZ, TLS, AND, SWZ, MNE, SLE, BDI, BHR, BIH, SUR, MAR, FJI, AZE, MDG, WSM, RWA, MUS, MOZ, NPL, ZAF, CIV, PAK, SAU, MWI, ECU, POL, COD, BFA, TGO, GUY, CAN, RUS, MNG, PER, SEN, BEN, TZA, BRN, CHE, NER, UGA, MLI, TUN, GNB, BRA, GMB, IND, SWE, TON, GBR, DNK, GHA, CMR, KWT, SLB, BLR, ROU, DZA, COL, COG. - Figure A.16: - Left panel compares bilateral geopolitical distance to the US and China in 2023; right panel compares bilateral geopolitical distance to the euro area and China in 2023. - Axis labels and scales preserved: Geopolitical distance to US with ticks 0 1 2 3 4; Geopolitical distance to CHN with ticks 0 1 2 3 4; Geopolitical distance to EA with ticks 0.5 1 1.5 2 2.5. - Country labels plotted in both panels include (selected examples exactly as shown): AGO, ALB, AND, ARG, ARM, AUS, AUT, AZE, BDI, BEL, BEN, BFA, BGD, BGR, BHR, BIH, BLR, BLZ, BRA, BRN, BWA, CAN, CHE, CHL, CIV, CMR, COD, COG, COL, CRI, CYP, CZE, DEU, DNK, DZA, ECU, EGY, ESP, EST, FIN, FJI, FRA, GBR, GEO, GHA, GMB, GNB, GRC, GUY, HRV, HUN, IDN, IND, IRL, ISL, ISR, ITA, JOR, JPN, KAZ, KEN, KGZ, KHM, KOR, KWT, LBR, LTU, LUX, LVA, MAR, MDA, MDG, MDV, MKD, MLI, MLT, MNE, MNG, MOZ, MUS, MWI, MYS, NAM, NER, NLD, NOR, NPL, NZL, PAK, PER, PHL, PNG, POL, PRT, PRY, ROU, RUS, RWA, SAU, SEN, SLB, SLE, SRB, SUR, SVK, SVN, SWE, SWZ, SYC, TGO, THA, TLS, TON, TUN, TUR, TZA, UGA, UKR, URY, UZB, WSM, ZAF, ZMB, ZWE. ### Geopolitics and invoicing currency regressions (Figures A.15 and A.17) - Figure A.15 (Panel regressions for export invoicing shares and geopolitical distance to the invoicing currency issuer post-2012): - Presents the coefficient estimate pγc,ipcq from Equation (3) for different choices of ̄tinIptą ̄tq. - Dots represent point estimates and whiskers indicate 90% confidence bands. - Rows: - First row: Dependent variable: US dollar invoicing shares. - Second row: Dependent variable: Euro invoicing shares. - Third row: Dependent variable: Renminbi invoicing shares. - Columns/variants shown include "Dummy non−zero from t>", "Commodity share as control", and "Oil share as control". - Year markers shown in the plots: 2013 2015 2017 2019 2021. - Inference is based on Driscoll-Kraay standard errors. - Axis ranges preserved in plots: for US dollar and euro panels coefficient ranges approximately −1 to 1 (and −1.5 to .5 in one euro panel); for renminbi panel coefficient ranges approximately −.15 to .1. - Figure A.17 (Geopolitics and export invoicing currency, pooled economy-currency panel regressions): - Presents qualitatively the coefficient estimates pγc,i from pooled economy-currency panel regressions analogous to Equation (4). - Rows indicate invoicing currency issuer (US, EA, CH); columns indicate invoicing currency (USD, EUR, CNY, HOME, OTHER). - A red cell indicates use of the corresponding invoicing currency is (conditionally) negatively correlated with a geopolitical distancing away from the corresponding invoicing currency issuer. - Only estimates statistically significant at the 10% level are displayed. - Left panel uses commodity export share in wce,t; right panel uses oil export share in wce,t. ### Regression tables: compositional/mechanical correlates and geopolitical distance (Tables B.1–B.8, B.1–B.8 highlights) - Table B.1: Panel regressions for compositional/mechanical correlates of import invoicing shares (dependent variables USD, EUR, CNY): - Lagged invoicing currency share coefficients: 0.82, 0.81, 0.84, 0.84, 0.94, 0.97 with standard errors shown as (0.00) for each. - Bilateral trade share with invoicing currency issuer coefficients: 0.05, 0.09, 0.11, 0.14, 0.01, 0.02 with p-values/SEs (0.31)(0.03)(0.00)(0.00)(0.00)(0.00). - Trade share with rest of invoicing currency block coefficients: 0.03, 0.04, 0.05, 0.13, −0.02, −0.01 with p-values/SEs (0.38)(0.22)(0.06)(0.00)(0.10)(0.14). - Bilateral exchange rate against invoicing currency coefficients reported as −0.00, 0.00 etc., with p-values/SEs (0.35)(0.99)(0.24)(0.14)(0.01)(0.01). - Commodity trade share: 0.05 (p-value/SE (0.03)); Oil trade share: 0.14 (p-value/SE (0.01)). - Year FEs: ✓ ; Within R-squared reported as 0.75, 0.75, 0.81, 0.82, 0.90, 0.92. - Observations: 134313, 271368, 134880, 8793 (numbers listed across columns exactly as shown); Countries: 113, 110, 113, 111, 96, 92. - Table B.2: Pooled economy-invoicing currency regressions for compositional/mechanical correlates of export invoicing shares: - Lagged invoicing currency share: 0.76 and 0.78 (p-values/SEs (0.00)(0.00)). - Bilateral trade share with invoicing currency issuer: 0.08 and 0.07 (p-values/SEs (0.00)(0.00)). - Trade share with rest of invoicing currency block: 0.01 and 0.04 (p-values/SEs (0.54)(0.02)). - Bilateral exchange rate: −0.00 and −0.00 (p-values/SEs (0.15)(0.46)). - Commodity trade share: −0.04 (p-value/SE (0.01)); Commodity trade shareˆIpc“USDq: 0.11 (p-value/SE (0.00)). - Oil trade share: −0.09 (p-value/SE (0.00)); Oil trade shareˆIpc“USDq: 0.25 (p-value/SE (0.00)). - Within R-squared: 0.71 and 0.71; Observations: 3316 and 3309; Country-currency pairs: 316 and 308. - Table B.3: Panel regressions for import invoicing shares and geopolitical distance (US, DE, EUR, CNY columns): - Lagged invoicing currency share repeated: 0.82, 0.81, 0.84, 0.83, 0.94, 0.97 (p-values/SEs (0.00) each). - Commodity trade share: 0.05 (p-value/SE (0.05)) and −0.00, −0.00 (p-values (0.98)(0.45)). - Oil trade share: 0.14 (p-value/SE (0.01)) and −0.05, 0.00 (p-values (0.15)(0.63)). - Geopolitical distance to invoicing currency issuer coefficients: 1.67, 1.05, −1.75, −2.36, 0.17, 0.10 with p-values/SEs (0.14)(0.41)(0.11)(0.00)(0.09)(0.18). - Geopolitical distance to invoicing currency issuerˆpost-2021 coefficients: −0.62, −0.91, −0.26, −0.16, −0.24, −0.21 with p-values/SEs (0.05)(0.01)(0.42)(0.53)(0.01)(0.00). - Standard controls: ✓ ; Year FEs: ✓ ; Within R-squared: 0.75, 0.75, 0.81, 0.82, 0.91, 0.92. - Observations: 134313, 271368, 134880, 6793; Countries: 113, 110, 113, 111, 95, 92. - p-value full effect reported per column: 0.25, 0.89, 0.03, 0.000, 0.29, 0.07. - Table B.4: Export invoicing shares and geopolitical distance excluding Russia, Belarus and Uzbekistan: - Commodity trade share coefficients: 0.09, −0.03, 0.00 (p-values (0.00)(0.06)(0.42)). - Oil trade share coefficients: 0.16, −0.16, −0.01 (p-values (0.01)(0.00)(0.06)). - Geopolitical distance to invoicing currency issuer coefficients: 1.11, 0.62, 0.10, −0.03, −0.08, −0.05 (p-values (0.15)(0.28)(0.83)(0.94)(0.02)(0.23)). - Geopolitical distance to invoicing currency issuerˆpost-2021 coefficients: −0.14, −0.10, −0.40, −0.63, −0.05, −0.03 (p-values (0.65)(0.63)(0.07)(0.00)(0.14)(0.11)). - Standard controls: ✓ ; Year FEs: ✓ ; Within R-squared: 0.64, 0.64, 0.77, 0.78, 0.83, 0.96. - Observations: 126112, 861266, 128369, 6704; Countries: 111, 109, 109, 108, 89, 88. - p-value full effect reported: 0.13, 0.36, 0.55, 0.14, 0.01, 0.04. - Table B.5: Export invoicing shares and alternative definitions of geopolitical distance to the euro area: - Columns labeled Baseline, Median, DE&FR, Trade-weighted with (1)-(8). - Commodity trade share coefficients: −0.04, −0.00, −0.00, −0.00 (p-values (0.04)(0.41)(0.42)(0.41)). - Oil trade share coefficients: −0.17 across multiple specifications (p-values (0.00) repeatedly). - Geopolitical distance to invoicing currency issuer coefficients and geopolitical distanceˆpost-2021 coefficients reported with several values; e.g., geopolitical distanceˆpost-2021 reported values such as −0.44, −0.77, −0.48, −0.78, −0.40, −0.58, −0.44, −0.69 with p-values/SEs (0.06)(0.00)(0.06)(0.00)(0.06)(0.00)(0.06)(0.00). - Standard controls and Year FEs: ✓ across columns. - Within R-squared: 0.77, 0.78, 0.78, 0.78, 0.78. - Observations reported exactly: 12841303, 12841303, 12841303 etc. (as shown). - Countries: 111, 110, 111, 110, 111, 110, 111, 110. - p-value full effect reported per column: 0.52, 0.08, 0.000, 0.13, 0.000, 0.000, 0.000, 0.01. - Tables B.6 and B.7: Pooled economy-currency regressions for export invoicing shares and geopolitical distance (including heterogeneous effects): - Table B.6 reports: - Geopolitical distance to invoicing currency issuer coefficients: 0.34 and −0.08 (p-values (0.23)(0.74)). - Geopolitical distance to invoicing currency issuer x post-2021: −0.08 and −0.27 (p-values (0.42)(0.00)). - Geopolitical distance x Ipc“$q: 0.89 and −0.39 (p-values (0.21)(0.62)). - Geopolitical distance x Ipc“$q x post-2021: −0.01 and −0.36 (p-values (0.98)(0.05)). - Geopolitical distance x Ipc“eq: −0.00 and −0.06 (p-values (0.99)(0.92)). - Geopolitical distance x Ipc“eq x post-2021: −0.31 and −0.49 (p-values (0.19)(0.02)). - Geopolitical distance x Ipc“¥q: −0.01 and −0.01 (p-values (0.91)(0.95)). - Geopolitical distance x Ipc“¥q x post-2021: 0.07 and 0.11 (p-values (0.20)(0.01)). - Within R-squared: 0.71 across columns; Observations: 3232, 3232, 3232, 3232 (and 3283 variants); Country-currency pairs: 311, 311, 306, 306. - Table B.7 reports similar specification with heterogeneous coefficients for trade shares, bilateral exchange rates and commodity/oil export shares; coefficients for geopolitical distance and interactions listed with p-values as shown (e.g., geopolitical distance to invoicing currency issuer 0.34 (0.22), −0.10 (0.70); interactions with post-2021 and Ipc indicators with p-values reported). - Table B.8: Correlation between US dollar invoicing shares and commodity export shares for US non-aligned countries post-2021: - Commodity trade share coefficients: 0.09, 0.10, 0.10 with p-values/SEs (0.00) repeated. - Commodity trade shareˆleast-alignedˆpost-2021 coefficients: −0.04, −0.04, −0.03 (p-values (0.11)(0.14)(0.15)). - Trade to GDP coefficients: −0.03, −0.03 (p-values (0.14)(0.14)). - Standard controls: ✓ ; Full set of interactions: ✓ ; Year FEs: ✓ ; YearˆLeast-aligned FEs: −✓ in column (2) and ✓ in column (3). - Within R-squared: 0.65, 0.66, 0.65. - Observations: 1298, 1274, 1257. - Countries: 114, 109, 107. - Note: Column (3) excludes Belarus and Russia relative to column (2). ### Appendix C — Assessing the usefulness of China’s settlement currency as proxy for invoicing currency - Definitions (exact notation preserved): - Share of CNY in China’s export invoicing: - s_{x,CNY}^{CHN} = \sum_i X_{CNY}^{CHN,i} / \sum_i X_{CHN,i} . (Equation (C.1) as presented) - Share of CNY in the invoicing of global imports from China: - s_{m,CNY}^{RoW} = \sum_i M_{CNY}^{i,CHN} / \sum_i M_{i,CHN} . (Equation (C.2) as presented) - Under bilateral consistency X_{CNY}^{CHN,i} = M_{CNY}^{i,CHN} and X_{CHN,i} = M_{i,CHN}, we have s_{m,CNY}^{RoW} = s_{x,CNY}^{CHN}. - Proxy condition: - The share of CNY in China’s cross-border export settlement, σ_{x,CNY}^{CHN}, is a useful proxy for s_{x,CNY}^{CHN} if σ_{x,CHN}^{CHN} « s_{m,CNY}^{RoW}. - Data complications noted exactly: - First complication: lack of information on the share of CNY invoicing in all countries’ imports; cannot directly calculate global CNY-invoiced imports from China without assumptions for countries not in dataset and country-years with incomplete currency information. - Definition of sets A (countries with CNY import invoicing information in dataset) and N (countries without such information) and resulting decomposition: - s_{m,CNY}^{RoW} = [\sum_{i \in A} M_{CNY}^{i,CHN} / \sum_i M_{i,CHN}] + [\sum_{i \in N} M_{CNY}^{i,CHN} / \sum_i M_{i,CHN}] = ω_{m}^{CHN,A} s_{m,CNY}^{A} + ω_{m}^{CHN,N} s_{m,CNY}^{N}. (Equation (C.3) as presented) - Given inability to calculate s_{m,CNY}^{N}, the implied s_{m,CNY}^{N} that would be required for σ_{x,CNY}^{CHN} to be a useful proxy is: - s_{m,CNY}^{N} = [σ_{x,CHN}^{CHN} − ω_{m}^{CHN,A} s_{m,CNY}^{A}] / ω_{m}^{CHN,N}. (Equation (C.4) as presented) - Second complication: dataset only contains information on countries’ total imports M_{CNY}^{i} but not on bilateral imports from China M_{CNY}^{i,CHN}. - The authors therefore make the assumption that CNY invoicing occurs only in bilateral trade with China, i.e., M_{CNY}^{i,j} = 0 for j ≠ CHN and hence M_{CNY}^{i,CHN} = s_{m,CNY}^{i} M_{i}. - This assumption biases in favor of China’s trade settlement currency being a useful proxy for trade invoicing currency by allocating the maximum possible amount of CNY invoicing of countries’ imports to imports from China. - Empirical note: - The green dashed line in the left panel in Figure A.2 presents results for the implied value of the CNY import invoicing shares s_{m,CHN}^{N} under the assumptions described. _Excerpted content as provided from the PDF chapter/section of Working Paper No. WP/2025/178._ --- _Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025178-source-pdf.pdf_