## Annex I. Additional Regression Results

## Source details

**Canonical URL:** [Annex I. Additional Regression Results](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024189-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024189-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024189-print-pdf.pdf.json)

---

### A. Context and objectives
- Geopolitical tensions have risen sharply: the Geopolitical Risk Index has "doubled in recent years."
- Trade restrictions and investment barriers have surged, pushing the Trade Policy Uncertainty Index to "a historic high."
- Paper objective: examine the effect of geopolitical proximity on the use of global currencies in cross-border transactions, focusing on the five SDR currencies (U.S. dollar, euro, Chinese RMB, Japanese yen, British pound).
- Empirical contribution: analyzes how geopolitical proximity affects currency usage across FX reserves, payment currencies, and invoicing, and how effects vary with overall global geopolitical tensions.

### B. Methodology and data (key specifications)
- Dependent variable: share of SWIFT flows (sum of flows sent and received) in reserve currency G over total flows (sent and received) between country s and country r in year t.
- Baseline panel specification includes a lagged dependent variable and main regressors:
  - Geopolitical proximity measured by correlation of UN voting outcomes with currency issuer G.
  - Trade and financial linkages (bilateral shares in trade, FDI and portfolio flows).
  - Geographical distance from the reserve currency issuer.
  - Legal tender dummy (equals 1 if currency G is legal tender in either sender or receiver).
  - Control vector X̄_sr: average of GDP, GDP per capita, financial development index, governance across the two transacting countries; year fixed effects included.
- RMB-specific dummies:
  - Bilateral swap arrangement with China (dummy).
  - Offshore RMB clearing bank (dummy).
- Non-linear specification: interaction between geopolitical proximity and two alternative indicators of global geopolitical tensions:
  - Index of international military conflicts (IMC).
  - Trade policy uncertainty index (TPI).
- Sample: 125 economies, annual data 2013–2021.
- Currency coverage: SDR basket currencies — U.S. dollar, euro, Japanese yen, British pound, Chinese renminbi; these five currencies together account for more than 85 percent of global cross-border transactions via SWIFT.
- Long-run effects formula: β2_LR = β2 / (1 − β1), where β2 is coefficient on geopolitical proximity and β1 is coefficient on lagged dependent variable.

### C. Key empirical findings — geopolitical proximity and currency shares
- Full sample:
  - Geopolitical proximity coefficient positive and statistically significant for RMB; coefficient for U.S. dollar is negative and statistically significant.
  - Non-intuitive U.S. dollar result: negative coefficient likely reflects that many U.S. allies are advanced economies with credible domestic currencies (e.g., AUD, CAD), reducing need to use USD; geopolitically distant EMDEs rely more on USD.
  - Excluding advanced economies makes the U.S. dollar geopolitical proximity coefficient close to zero and no longer statistically significant.
- EMDE sample:
  - Geopolitical proximity has a more substantial impact on currency choice:
    - Euro: a one-percentage point (ppt) increase in voting correlations with the Euro Area is likely to raise the euro share in cross-border payments by 0.19 ppt.
    - RMB: coefficient 0.02 ppt (statistically significant in EMDE sample).
    - British pound: positive and significant coefficient of 0.008 (smaller quantitative impact).
    - Japanese yen and U.S. dollar: statistically insignificant in EMDE sample.
  - Long-run effects (EMDE example using reported lag coefficients):
    - Euro long-run effect: 0.19 / (1 − 0.72) = 0.7 ppt.
    - RMB long-run effect: 0.02 / (1 − 0.84) = 0.13 ppt.
  - Example transition: moving from a voting correlation of 0.5 (50 ppt) to almost 1 (100 ppt) could boost the share of the euro and the RMB by 35 ppt and 8.5 ppt, respectively.

### D. Non-linear effects (interactions with global tensions)
- Interaction with international military conflicts (IMC):
  - Euro usage effect increases from 0.17 ppt to 0.23 ppt (increase of almost 40 percent) following a one-standard-deviation increase in IMC (Table B2).
  - RMB impact is less sensitive to IMC.
- Interaction with trade policy uncertainty (TPI):
  - RMB impact increases from 0.008 ppt to 0.013 ppt (increase of 64 percent) following a one-standard-deviation increase in TPI (Table A1).
  - For GBP, JPY, and USD the non-linear impact with TPI is not statistically significant.
- Interpretation: heightened global tensions are associated with diversification away from dominant currencies to enhance security and resilience of cross-border transactions.

### E. Other drivers of currency usage (estimated effects)
- Trade linkages:
  - Full sample: a one-ppt increase in trade share boosts euro usage by 0.08 ppt and U.S. dollar usage by 0.1 ppt.
  - EMDEs: impacts larger — 0.25 ppt for euro and 0.14 ppt for U.S. dollar.
  - RMB: trade linkage effect weaker at 0.02 ppt.
- Financial linkages:
  - FDI share: one-ppt increase in FDI share with China and the U.K. boosts use of RMB and GBP by 0.02 ppt.
  - Portfolio flows share: statistically significant impacts across most currencies:
    - U.S. dollar: 0.05
    - Euro: 0.04
    - RMB: 0.09
  - Stronger portfolio flow impact for RMB reflects opening of China’s local currency bond market and foreign inflows.
  - GBP: portfolio flows coefficient statistically insignificant in full sample regressions.
- Legal tender:
  - Coefficients in the range of 4–17 (Table 1), implying long-run increases in transaction shares of 22–72 percentage points if the major currency is legal tender in one or both countries.
- Geographical distance:
  - Tends to dampen euro usage; effects vary by currency.
- Financial development:
  - More financially developed economies tend to have lower shares of transactions in euro and U.S. dollar, consistent with advanced economies using local currencies.

### F. Robustness checks and specification notes
- Linkage variables constructed as averages across sender and receiver: Z̄_sr,t ≡ (Z_s,t + Z_r,t) / 2; alternative GDP-weighted averages tested and baseline results robust.
- Euro area treatment:
  - When G = euro, geopolitical proximity and geographical distance use the average for France and Germany; trade and financial linkages use the sum of all Euro Area countries vis-à-vis transacting countries.
  - Bilateral observations within the Euro Area are excluded.
- RMB robustness:
  - Excluding transactions with Hong Kong SAR, Macau SAR, and Taiwan POC yields quantitatively similar results.
- Country fixed effects omitted due to little time variation in geopolitical proximity and potential Nickell (1981) bias with lagged dependent variable.
- Sample and period: 125 economies, yearly 2013–2021; SWIFT data use MT 103/202 message types and net USD amounts.

### G. Regression tables — selected coefficient summaries (as reported)
- Full-sample interaction with Trade Policy Uncertainty (columns: EUR GBR JPY RMB USD):
  - Currency Share (-1): 0.806***0.700***0.765***0.873***0.823***
  - UN Voting Proximity: 0.0130.0030.0010.008*-0.134***
  - UN Voting Proximity * TPI: 0.013**-0.0003-0.002***0.005**0.002
  - Trade Share: 0.076**0.018-0.0090.022**0.096***
  - FDI Share: 0.0020.023*0.0030.020**-0.012
  - Portfolio Flows Share: 0.036**-0.005*0.003***0.089**0.053***
  - GDP (log): 0.318***-0.064***-0.013-0.068*-0.183
  - GDP per Capita (log): 0.1060.061**0.0390.182***-0.545
  - Financial Development Index: -9.693***0.0360.1850.516**-5.092**
  - Geographical Distance (log): -1.089***-0.033-0.0900.327*1.173**
  - Legal Tender: 4.511***7.052***16.96***6.096***7.254***
  - AML: -0.0900.0190.049**0.0280.025
  - Chinese BSL Dummy: 0.041
  - Offshore RMB Bank Dummy: 0.070*
  - Constant: YesYesYesYesYes
  - Time Fixed Effects: YesYesYesYesYes
  - Observation Number: 7901083838724998009983667
  - R-squared: 0.740.570.900.770.74
- Full-sample interaction with IMC (columns: EUR GBR JPY RMB USD):
  - Currency Share (-1): 0.806***0.700***0.765***0.873***0.823***
  - UN Voting Proximity: 0.0090.0020.0010.009-0.129***
  - UN Voting Proximity * IMC: 0.0180.006**0.001-0.001-0.018
  - Trade Share: 0.077**0.019-0.0090.022**0.095***
  - FDI Share: 0.0020.022*0.0030.020**-0.012
  - Portfolio Flows Share: 0.036**-0.005*0.003***0.088**0.053***
  - GDP (log): 0.314***-0.063***-0.013-0.068*-0.183
  - GDP per Capita (log): 0.1050.058**0.0380.181***-0.540
  - Financial Development Index: -9.642***0.0480.1860.516**-5.146**
  - Geographical Distance (log): -1.089***-0.033-0.0900.328*1.181**
  - Legal Tender: 4.509***7.052***16.96***6.097***7.260***
  - AML: -0.0870.0180.048**0.0270.022
  - Intra-Euro Area Dummy: 0.000
  - Chinese BSL Dummy: 0.040
  - Offshore RMB Bank Dummy: 0.076*
  - Constant: YesYesYesYesYes
  - Time Fixed Effects: YesYesYesYesYes
  - Observation Number: 7901083838724998009983667
  - R-squared: 0.740.570.900.770.74
- EMDE-sample interaction with Trade Policy Uncertainty (columns: EUR GBR JPY RMB USD):
  - Currency Share (-1): 0.719***0.556***0.1670.842***0.816***
  - UN Voting Proximity: 0.184***0.008*-0.0010.019***-0.057
  - UN Voting Proximity * TPI: 0.0290.0030.00030.00005-0.010
  - Trade Share: 0.251***-0.0090.020**0.0090.136**
  - FDI Share: 0.0010.023*-0.0020.0020.044
  - Portfolio Flows Share: 0.005-0.001-0.003**-0.0060.039**
  - GDP (log): -0.290*-0.030*-0.023*0.003-0.053
  - GDP per Capita (log): -0.0030.295***-0.0210.087-1.752***
  - Financial Development Index: -0.827-0.5260.260-0.807***-1.979
  - Geographical Distance (log): -1.938***0.161-0.0500.0691.837
  - Legal Tender: 8.558***3.991***
  - AML: -0.1780.00030.050**-0.0430.376
  - Chinese BSL Dummy: 0.143
  - Offshore RMB Bank Dummy: 0.013
  - Constant: YesYesYesYesYes
  - Time Fixed Effects: YesYesYesYesYes
  - Observation Number: 2095121304158131946321293
  - R-squared: 0.660.270.030.790.69
- EMDE-sample interaction with IMC (columns: EUR GBR JPY RMB USD):
  - Currency Share (-1): 0.719***0.556***0.1670.842***0.816***
  - UN Voting Proximity: 0.167***0.008*-0.0010.019***-0.043
  - UN Voting Proximity * IMC: 0.066*-0.00040.0010.001-0.051
  - Trade Share: 0.252***-0.0090.020**0.0090.136**
  - FDI Share: 0.0020.024*-0.0020.0020.045
  - Portfolio Flows Share: 0.006-0.001-0.003**-0.0060.040**
  - GDP (log): -0.293*-0.031*-0.023*0.003-0.053
  - GDP per Capita (log): -0.0320.297***-0.0220.087-1.714***
  - Financial Development Index: -0.720-0.5270.262-0.806***-2.054
  - Geographical Distance (log): -1.929***0.160-0.0510.0701.855
  - Legal Tender: 8.558***3.997***
  - AML: -0.2200.0030.049**-0.0420.417
  - Chinese BSL Dummy: 0.143
  - Offshore RMB Bank Dummy: 0.013
  - Constant: YesYesYesYesYes
  - Time Fixed Effects: YesYesYesYesYes
  - Observation Number: 2095121304158131946321293
  - R-squared: 0.660.270.030.790.69

### H. Data appendix highlights (data sources and construction)
- SWIFT data:
  - Dependent variable: share of currency usage between country pairs for the five SDR basket currencies.
  - SWIFT message types: MT 103 (MT103, MT103+, MT103R) and MT 202/MT 202C.
  - Panel frequency: yearly 2013–2021; currency shares calculated based on SWIFT net amount sent and received in USD.
- Geopolitical proximity:
  - Based on UN General Assembly voting data (Voeten et al., 2009); Arend Lijphart’s Voting Similarity Index formula reported with index range between zero and one.
- Measures of geopolitical tensions:
  - International Military Conflict Index: number of total interstate and internationalized intrastate conflicts (Uppsala Conflict Data Program).
  - Trade Policy Uncertainty Index: frequency of joint occurrences of trade policy and uncertainty terms across major newspapers.
- Country-pair macro variables and country-level data sources listed (IMF DOTS, CPIS, CDIS, WEO, IMF Financial Development Index, AREAER, Basel AML, GeoDist, People’s Bank of China for Chinese BSL and offshore clearing data).

### I. Policy implications and conclusions (concise)
- Closer geopolitical proximity can boost the use of alternative reserve currencies (euro and RMB), notably among EMDEs; effects are muted in full sample.
- RMB usage effect in the full sample is more pronounced during periods of heightened trade policy uncertainty.
- Trade and financial ties with currency issuers are important drivers:
  - Trade linkages strong for euro and USD; weaker for RMB.
  - Portfolio flows important, particularly for RMB.
- High inertia in currency usage across SDR currencies implies past transaction patterns strongly influence current preferences.
- Legal tender status significantly enhances currency usage in bilateral transactions.
- Given correlation between payment currency and reserve currency configuration, shifts at the payment level could be reflected in reserve configurations with lags due to market liquidity constraints.
- In a more geoeconomically fragmented world, alternative currencies could play a greater role in cross-border transactions; transition speed uncertain and could involve financial volatility, underscoring the importance of international coordination to mitigate adverse effects on the IMS and global stability.

*Annex I and Annex II, "Additional Regression Results" and "Data Appendix" from Geopolitics and the Use of Global Currencies, Working Paper No. WP/2024/189*

### Annex I. Additional Regression Results .................................................................................

### Annex I. Additional Regression Results

### I. Introduction — Context and motivation
- Geopolitical tensions have risen sharply: the Geopolitical Risk Index has "doubled in recent years."
- Trade restrictions and investment barriers have surged, pushing the Trade Policy Uncertainty Index to "a historic high."
- Key potential impacts of geoeconomic fragmentation:
  - Channels: trade, labor, capital, technology, and the provision of global public goods.
  - Bolhuis et al. (2023) estimate trade fragmentation could reduce global output by "0.2–7 percent."
  - IMF (2023a, 2023b) show geopolitical distance (correlation in UN General Assembly voting outcomes) influences FDI, portfolio, and banking flows.
- Paper objective: examine the effect of geopolitical proximity on the use of global currencies in cross-border transactions, focusing on the five SDR currencies (U.S. dollar, euro, Chinese RMB, Japanese yen, British pound).
- Empirical contribution: analyzes how geopolitical proximity affects currency usage across FX reserves, payment currencies, and invoicing, and how effects vary with overall global geopolitical tensions.

### II. Historical perspective of the International Monetary System (IMS)
- Pre-World War I: gold standard underpinned the IMS; pound sterling dominated international trade.
- Transition phases:
  - Gold exchange standard: U.S. dollar and pound sterling shared reserve roles.
  - Bretton Woods (post-World War II): U.S. dollar pegged to gold at "35 dollars per ounce"; other currencies pegged to the U.S. dollar.
  - 1969: IMF introduced Special Drawing Rights (SDRs) as a supplement to reserve assets.
  - 1971: President Nixon suspended the gold window; G10 signed the Smithsonian Agreement.
  - 1973: Bretton Woods collapsed; era of floating exchange rates and increasing capital account liberalization began.
- Despite large shifts in global economic weight, the IMS has shown strong inertia:
  - EMDEs now account for "close to 60 percent of global GDP (in PPP terms), up from less than 40 percent in the early 1990s."
  - EMDE nominal share rose "from 20 percent to more than 40 percent."
  - Aggregate international currency usage: U.S. dollar "continues to serve as the dominant currency, accounting for more than 60 percent of global usage," euro "around 20 percent."
- Historical episodes of rapid change: inter-war period saw volatile reserve currency shares and the IMS splitting into multiple blocs by 1932.

### III. Global currency configuration — functions and observed patterns
- Functions of money and corresponding global configurations:
  - Unit of account: trade invoicing, international debt issuance, FX turnover.
  - Medium of exchange: global payment currency.
  - Store of value: foreign exchange reserves.
- Complementarity across functions: currencies with stable value are more likely to serve as payment media and units of account.
- Concentration of international currency use:
  - Although >150 legal tender currencies exist, only a small number function extensively as international currencies.
  - Snapshot indicators (notes from Figure 6): FX reserves, international debt, international loans (latest data for these are "fourth quarter of 2022"); SWIFT data are for "December 2022"; FX turnover data are as of "April 2022".
  - In FX-market shares, since transactions involve two currencies, shares add up to "200%."
- Regional variation in payment currency usage (Figure 7):
  - U.S. dollar accounts for "more than half of payments in most regions."
  - Euro plays an eminent role in most of Europe and parts of Africa.
  - RMB has gained traction in parts of Asia (e.g., Mongolia and Laos).
  - Notably, the U.S. dollar has a larger presence in China for cross-border payments than the RMB itself.
  - Yen outside Japan is mainly present in Thailand’s cross-border transactions.
  - British pound used frequently in Europe and parts of Africa.

### IV. Geopolitical proximity and currency configuration — measurement and patterns
- Standard measure: correlation between countries' votes at the United Nations General Assembly (higher correlation = closer political proximity).
- Political proximity to reserve currency issuers (2022) — descriptive patterns (Figure 8):
  - U.S.: closely aligned with Australia, Canada, the U.K., and continental Europe.
  - European countries: broader geopolitical alignment spanning parts of Asia, the Middle East, Latin America, North America, Australia, and New Zealand.
  - China: close geopolitical proximity with most developing countries, including Asian, African, and Latin American countries.
- Research gap addressed: while prior studies analyze geopolitical proximity between transacting parties or political proximity to China, this paper provides a comprehensive analysis of political proximity to the five SDR issuers and how such proximity affects the use of each major currency in cross-border payments, including variation during periods of heightened geopolitical tensions.

*Source: IMF Working Paper — Geoeconomic Proximity and Global Currency Configuration (excerpts from the PDF content unit).*

### 4.2 Impact of Geopolitical Proximity

### 4.2 Impact of Geopolitical Proximity

### Methodology and data
- Dependent variable: share of SWIFT flows (sum of flows sent and received) in reserve currency G over total flows (sent and received) between country s and country r in year t.
- Baseline panel specification includes:
  - Lagged dependent variable to capture inertia in currency usage.
  - Main regressors: geopolitical proximity (measured by the correlation of UN voting outcomes of country s and r with the global currency issuer G), trade and financial linkages (bilateral shares in trade, FDI and portfolio flows), geographical distance from the reserve currency issuer, and legal tender dummy (equals 1 if currency G is legal tender in either sender or receiver).
  - Control vector X̄_sr capturing average values across the two transacting countries: GDP, GDP per capita, financial development index, governance, plus year fixed effects.
- For RMB additional dummies:
  - i) equals 1 if either sender or receiver has a bilateral swap arrangement with China.
  - ii) equals 1 if either sender or receiver has an offshore RMB clearing bank.
- Non-linear specification: interaction between geopolitical proximity and measures of global geopolitical tensions (two alternative indicators):
  - Index of international military conflicts (higher = more military confrontations).
  - Trade policy uncertainty index (frequency of mentions of trade policy and uncertainty terms in major newspapers).
- Sample: 125 economies, annual data spanning from 2013 to 2021.
- Currency coverage: SDR basket currencies — U.S. dollar, euro, Japanese yen, British pound, and Chinese renminbi. These five currencies together account for more than 85 percent of global cross-border transactions via SWIFT.
- Long-run effects computed as β2_LR = β2 / (1 − β1), where β2 is the coefficient on geopolitical proximity and β1 is the coefficient on the lagged dependent variable.

### Key empirical findings
- Geopolitical proximity effect (full sample and EMDEs):
  - Full sample: geopolitical proximity coefficient positive and statistically significant for RMB; coefficient for U.S. dollar is negative and statistically significant.
  - EMDE sample: geopolitical proximity has a more substantial impact on currency choice.
    - Euro: a one-percentage point (ppt) increase in voting correlations with the Euro Area is likely to raise the euro share in cross-border payments by 0.19 ppt.
    - RMB: the impact is statistically significant with coefficient 0.02 ppt (EMDE sample).
    - British pound: positive and significant coefficient of 0.008 (quantitative impact smaller).
    - Japanese yen and U.S. dollar: statistically insignificant in EMDE sample.
  - Long-run effects (EMDE example using reported lag coefficients):
    - Euro long-run effect: 0.19 / (1 − 0.72) = 0.7 ppt.
    - RMB long-run effect: 0.02 / (1 − 0.84) = 0.13 ppt.
  - Example transition: moving from a voting correlation of 0.5 (50 ppt) to almost 1 (100 ppt) could boost the share of the euro and the RMB by 35 ppt and 8.5 ppt, respectively.
- Non-intuitive U.S. dollar result in full sample:
  - Negative coefficient for geopolitically closer alignment with the U.S. possibly reflects that many U.S. allies are advanced economies with strong domestic currencies (e.g., Australian dollar, Canadian dollar) and hence less need to use the U.S. dollar in payments.
  - Geopolitically distant countries are often EMDEs lacking credible domestic currencies and therefore rely more on the U.S. dollar as a global vehicle currency.
  - Excluding advanced economies makes the U.S. dollar geopolitical proximity coefficient close to zero and no longer statistically significant.

### Non-linear effects of geopolitical proximity
- Interaction with international military conflicts:
  - Euro usage effect increases from 0.17 ppt to 0.23 ppt (increase of almost 40 percent) following a one-standard-deviation increase in international military conflicts (Table B2).
  - RMB impact is less sensitive to military conflicts.
- Interaction with trade policy uncertainty:
  - RMB impact increases from 0.008 ppt to 0.013 ppt (increase of 64 percent) following a one-standard-deviation increase in trade policy uncertainty (Table A1).
  - For British pound, Japanese yen, and U.S. dollar the non-linear impact is not statistically significant.
- Interpretation: when global tensions increase, countries may attempt to diversify away from dominant currencies to enhance security and resilience of cross-border transactions.

### Other drivers of currency usage
- Trade linkages:
  - Full sample: a one-ppt increase in trade share boosts euro usage by 0.08 ppt and U.S. dollar usage by 0.1 ppt.
  - EMDEs: the impact is larger — 0.25 ppt for euro and 0.14 ppt for U.S. dollar.
  - RMB: trade linkage effect is weaker at 0.02 ppt, reflecting limited use of RMB as an invoicing currency in global trade during the sample period.
- Financial linkages:
  - FDI share: one-ppt increase in FDI share with China and the U.K. boosts use of RMB and British pound by 0.02 ppt (small positive impact).
  - Portfolio flows share: statistically significant impact across most major currencies:
    - U.S. dollar: 0.05
    - Euro: 0.04
    - RMB: 0.09
  - Stronger portfolio flow impact for RMB likely reflects opening of China’s local currency bond market and subsequent foreign inflows.
  - British pound: portfolio flows statistically insignificant, possibly reflecting the U.K. role as a global financial center and offshore funding market dynamics.
- Legal tender:
  - Coefficients in the range of 4–17 (Table 1), implying in long-run terms that the share of the respective major currency used in transactions between two countries would increase by 22–72 percentage points if the major currency is a legal tender in one or both countries.
- Geographical distance:
  - Tends to dampen euro usage; impacts differ by currency.
- Financial development:
  - More financially developed economies tend to have lower shares of transactions in euro and U.S. dollar, consistent with advanced economies using their local currencies (e.g., Australian dollar, Swiss franc).

### Robustness and supplementary notes
- Construction of linkage variables:
  - To reduce regressors, trade, financial, and geopolitical linkage variables are constructed by averaging sender and receiver values: Z̄_sr,t ≡ (Z_s,t + Z_r,t) / 2.
  - Alternative weighting by relative GDP size of sender and receiver was tested and baseline results are robust.
- Euro area treatment:
  - When G = euro, geopolitical proximity and geographical distance use the average value for France and Germany; trade and financial linkages use the sum of all Euro Area countries vis-à-vis transacting countries.
  - Bilateral observations between countries within the Euro Area are excluded given euro usage within the area.
- RMB robustness:
  - Regressions excluding transactions with Hong Kong SAR, Macau SAR, and Taiwan POC yield quantitatively similar results.
- Omission of country fixed effects:
  - Country fixed effects not included because geopolitical proximity has little time variation and including country fixed effects with a lagged dependent variable could introduce bias (Nickell, 1981).

### Policy implications and conclusions
- Closer geopolitical proximity can boost the use of alternative reserve currencies (euro and RMB), notably among EMDEs, while effects are muted in the full sample.
- RMB usage effect in the full sample is more pronounced during periods of heightened trade policy uncertainty.
- Trade and financial ties with currency issuers are important drivers of currency usage:
  - Trade linkages are strong drivers for the euro and U.S. dollar, less so for RMB.
  - Portfolio flows are an important driver, particularly for RMB.
- High inertia in currency usage across all five SDR currencies implies past transaction patterns strongly influence current preferences.
- Legal tender status significantly enhances currency usage in transactions.
- Given correlation between payment currency and reserve currency configuration, geopolitical proximity impacts at the payment level could eventually be reflected in reserve currency configuration, albeit potentially lagged because of market liquidity constraints in alternative currencies relative to the U.S. dollar.
- In a more geoeconomically fragmented world, alternative currencies could play a greater role in cross-border transactions; the speed of transition is uncertain and could be accompanied by financial volatility, making international coordination important to mitigate adverse effects on the IMS and global economic stability.

*IMF Working Paper — Geoeconomic Proximity and Global Currency Configuration (Section 4.2)*

### Annex I.  Additional Regression Results

### Annex I.  Additional Regression Results

### A. Regression Results on Interaction with Global Geopolitical Tensions (Full Sample)

- Table A1. Regression on Results on Interaction with Trade Policy Uncertainty
  - Note: ***, **, and * indicate statistical significance at 1%, 5% and 10%, respectively, where the standard errors are clustered at the year level. Bolded numbers are numbers that are statistically significant at 10% level or lower.
- Table A2. Regression on Results on Interaction with International Military Conflicts
  - Note: ***, **, and * indicate statistical significance at 1%, 5% and 10%, respectively, where the standard errors are clustered at the year level. Bolded numbers are numbers that are statistically significant at 10% level or lower.

- Regression coefficient summaries (columns: EUR GBR JPY RMB USD as presented)
  - Currency Share (-1): 0.806***0.700***0.765***0.873***0.823***
  - UN Voting Proximity: 0.0130.0030.0010.008*-0.134***
  - UN Voting Proximity * TPI: 0.013**-0.0003-0.002***0.005**0.002
  - Trade Share: 0.076**0.018-0.0090.022**0.096***
  - FDI Share: 0.0020.023*0.0030.020**-0.012
  - Portfolio Flows Share: 0.036**-0.005*0.003***0.089**0.053***
  - GDP (log): 0.318***-0.064***-0.013-0.068*-0.183
  - GDP per Capita (log): 0.1060.061**0.0390.182***-0.545
  - Financial Development Index: -9.693***0.0360.1850.516**-5.092**
  - Geographical Distance (log): -1.089***-0.033-0.0900.327*1.173**
  - Legal Tender: 4.511***7.052***16.96***6.096***7.254***
  - AML: -0.0900.0190.049**0.0280.025
  - Chinese BSL Dummy: 0.041
  - Offshore RMB Bank Dummy: 0.070*
  - Constant: YesYesYesYesYes
  - Time Fixed Effects: YesYesYesYesYes
  - Observation Number: 7901083838724998009983667
  - R-squared: 0.740.570.900.770.74

- Regression coefficient summaries with IMC interaction (columns: EUR GBR JPY RMB USD as presented)
  - Currency Share (-1): 0.806***0.700***0.765***0.873***0.823***
  - UN Voting Proximity: 0.0090.0020.0010.009-0.129***
  - UN Voting Proximity * IMC: 0.0180.006**0.001-0.001-0.018
  - Trade Share: 0.077**0.019-0.0090.022**0.095***
  - FDI Share: 0.0020.022*0.0030.020**-0.012
  - Portfolio Flows Share: 0.036**-0.005*0.003***0.088**0.053***
  - GDP (log): 0.314***-0.063***-0.013-0.068*-0.183
  - GDP per Capita (log): 0.1050.058**0.0380.181***-0.540
  - Financial Development Index: -9.642***0.0480.1860.516**-5.146**
  - Geographical Distance (log): -1.089***-0.033-0.0900.328*1.181**
  - Legal Tender: 4.509***7.052***16.96***6.097***7.260***
  - AML: -0.0870.0180.048**0.0270.022
  - Intra-Euro Area Dummy: 0.000
  - Chinese BSL Dummy: 0.040
  - Offshore RMB Bank Dummy: 0.076*
  - Constant: YesYesYesYesYes
  - Time Fixed Effects: YesYesYesYesYes
  - Observation Number: 7901083838724998009983667
  - R-squared: 0.740.570.900.770.74

### B. Regression Results on Interaction with Global Geopolitical Tensions (EMDE Sample)

- Table B1. Regression Results on Interaction with Trade Policy Uncertainty
  - Note: ***, **, and * indicate statistical significance at 1%, 5% and 10%, respectively, where the standard errors are clustered at the year level. Bolded numbers are numbers that are statistically significant at 10% level or lower.
- Table B2. Regression Results on Interaction with International Military Conflicts
  - Note: ***, **, and * indicate statistical significance at 1%, 5% and 10%, respectively, where the standard errors are clustered at the year level. Bolded numbers are numbers that are statistically significant at 10% level or lower.

- EMDE sample regression coefficient summaries (columns: EUR GBR JPY RMB USD as presented)
  - Currency Share (-1): 0.719***0.556***0.1670.842***0.816***
  - UN Voting Proximity: 0.184***0.008*-0.0010.019***-0.057
  - UN Voting Proximity * TPI: 0.0290.0030.00030.00005-0.010
  - Trade Share: 0.251***-0.0090.020**0.0090.136**
  - FDI Share: 0.0010.023*-0.0020.0020.044
  - Portfolio Flows Share: 0.005-0.001-0.003**-0.0060.039**
  - GDP (log): -0.290*-0.030*-0.023*0.003-0.053
  - GDP per Capita (log): -0.0030.295***-0.0210.087-1.752***
  - Financial Development Index: -0.827-0.5260.260-0.807***-1.979
  - Geographical Distance (log): -1.938***0.161-0.0500.0691.837
  - Legal Tender: 8.558***3.991***
  - AML: -0.1780.00030.050**-0.0430.376
  - Chinese BSL Dummy: 0.143
  - Offshore RMB Bank Dummy: 0.013
  - Constant: YesYesYesYesYes
  - Time Fixed Effects: YesYesYesYesYes
  - Observation Number: 2095121304158131946321293
  - R-squared: 0.660.270.030.790.69

- EMDE sample regression coefficient summaries with IMC interaction (columns: EUR GBR JPY RMB USD as presented)
  - Currency Share (-1): 0.719***0.556***0.1670.842***0.816***
  - UN Voting Proximity: 0.167***0.008*-0.0010.019***-0.043
  - UN Voting Proximity * IMC: 0.066*-0.00040.0010.001-0.051
  - Trade Share: 0.252***-0.0090.020**0.0090.136**
  - FDI Share: 0.0020.024*-0.0020.0020.045
  - Portfolio Flows Share: 0.006-0.001-0.003**-0.0060.040**
  - GDP (log): -0.293*-0.031*-0.023*0.003-0.053
  - GDP per Capita (log): -0.0320.297***-0.0220.087-1.714***
  - Financial Development Index: -0.720-0.5270.262-0.806***-2.054
  - Geographical Distance (log): -1.929***0.160-0.0510.0701.855
  - Legal Tender: 8.558***3.997***
  - AML: -0.2200.0030.049**-0.0420.417
  - Chinese BSL Dummy: 0.143
  - Offshore RMB Bank Dummy: 0.013
  - Constant: YesYesYesYesYes
  - Time Fixed Effects: YesYesYesYesYes
  - Observation Number: 2095121304158131946321293
  - R-squared: 0.660.270.030.790.69

### Annex II. Data Appendix

- SWIFT Data
  - Dependent variable: share of currency usage between country pairs for the five SDR basket currencies.
  - SWIFT transfer types: single customer credit transfers (SWIFT message type MT 103, including MT103, MT103+, MT103R) and general financial institutions transfers (SWIFT message type MT 202 and MT 202C).
  - Panel frequency and period: yearly frequency from 2013 to 2021.
  - Currency shares for USD, RMB, EUR, JPY, and GBP calculated based on SWIFT net amount sent and received in USD; each observation corresponds to a unique sender country, receiver country, currency, and year.

- Geopolitical Proximity
  - Based on United Nations General Assembly voting data from Harvard Dataverse (Voeten et al., 2009).
  - Arend Lijphart’s Voting Similarity Index formula as presented:
    - (PP−1/2 G) G ×100  where PP is number of votes when both countries agree, G is number of votes when one country abstains and the other country votes “yes” or “no”, G is total number of votes.
    - Index range: between zero and one; one means identical voting behavior; closer to zero means divergent voting behavior.

- Measures of Geopolitical Tensions
  - International Military Conflict Index from Uppsala Conflict Data Program: represents number of total interstate and internationalized intrastate conflicts in a year. Definitions of interstate and internationalized conflicts provided.
  - Trade Policy Uncertainty Index: constructed by counting frequency of joint occurrences of trade policy and uncertainty terms across major newspapers.

- Country-pair Macro Variables (data sources listed)
  - Bilateral trade data: IMF Directions of Trade Statistics Database.
  - Bilateral portfolio investment positions: IMF Coordinated Portfolio Investment Survey Database.
  - Bilateral direct investment positions: IMF Coordinated Direct Investment Survey database.
  - Geographical distance data: GeoDist database (CEPII).

- Country Level Data (data sources listed)
  - GDP and per capita income: IMF World Economic Outlook (WEO) database.
  - Financial Development Index: IMF Financial Development Index Database (measures depth, access, efficiency).
  - Legal tender dummy: IMF AREAER Database.
  - Basel AML index: Basel Institute on Governance (captures risk of money laundering and terror financing).
  - Chinese Bilateral Swap Lines and RMB Offshore Clearing Bank data: People’s Bank of China.

*Annex I and Annex II, "Additional Regression Results" and "Data Appendix" from Geopolitics and the Use of Global Currencies, Working Paper No. WP/2024/189*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024189-print-pdf.pdf_