## 1. Currencies of Global Balance Sheets

## Source details

**Canonical URL:** [1. Currencies of Global Balance Sheets](https://www.imf.org/-/media/files/publications/esr/2025/english/ch2annex.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/esr/2025/english/ch2annex.pdf.md)
- [Structured JSON version](/-/media/files/publications/esr/2025/english/ch2annex.pdf.json)

---

### Method for Global Balance Sheet (GBS) index
- GBS index for currency i is defined as (퐴퐴_i + 퐿퐿_i) / (퐴퐴 + 퐿퐿), where 퐴퐴_i = foreign assets in currency i, 퐿퐿_i = foreign liabilities in currency i, and 퐴퐴 + 퐿퐿 = total foreign assets and liabilities across all currencies.
- For both assets and liabilities, estimated currency shares are used for each asset class (FDI, banking, debt, and reserves) multiplied by the share of that asset class in total foreign assets and liabilities.
- Currency weights by asset class are held constant at 2020 values (latest observation based on Allen and Juvenal, 2025) and multiplied by holdings using the latest vintage of foreign assets and liabilities data (Milesi-Ferretti, 2025 and extended with balance of payments data).
- For the euro area, intra-EU holdings are excluded.
- Currency weights are based on a confidential survey combined with publicly available data (see Allen and Juvenal (2025)).

### Outcome
- Figure 2.1.1 (not reproduced here) shows the resulting GBS index constructed as described.

---

### 2. Frankel and Wei Based Global Currency Blocs

### Methodology (two main steps)
- Step 1: Rolling monthly regressions of log returns of each currency s on log returns of global currencies (US dollar, euro, Japanese yen, British pound) expressed in the SDR numeraire:
  - Equation estimated: ∆log(s_GSSD)_t = α + Σ β_s t i ∆log(global_currency_i_GSSD)_t + ε_t.
  - The renminbi is excluded from estimation because it has been stable vis-à-vis the dollar in most sample years.
  - Following Ito and McCauley (2019) and Vicquéry (2022), all negative estimated coefficients are set to zero.
- Step 2: World divided into GDP-weighted currency zones according to annual co-movement of each economy’s currency with global currencies; economies can be apportioned to multiple global currencies in a given year.
  - Reserve-issuing bloc is assigned to its own currency bloc.
  - Currency weights for each economy are multiplied by GDP in US dollars and aggregated to the global economy.

### Outcome
- Figure 2.1.2 (not reproduced here) shows the resulting currency blocs over time.
- Methodology builds on Frankel and Wei (1994) and follows practices in Ito and McCauley (2019), Vicquéry (2022), and related literature.

---

### 3. Network Centrality Measures

### Eigenvector centrality used to assess importance in trade and financial networks
- Centrality score c_i for country i defined by: λ c_i = Σ g_j i c_j for all i ∈ N, where g denotes the adjacency matrix capturing connection strength.
- In directed networks, g = g_ji represents exports from country j to i or j’s holdings of i’s assets, expressed as a share of world GDP.
- In undirected networks, g = exports + imports between i and j, or sum of cross-holdings of financial assets and liabilities.
- λ is the largest eigenvector; centrality scores are normalized to sum to 1 (Σ c_j = 1).
- Interpretation: A country’s centrality is proportional to the sum of the centrality of its neighbors; linkages with more important countries receive higher weight.

---

### 4. Snapshot of the International Monetary System (selected statistics)

### 1 – Current State of International Currencies (percent of total; time indicated)
- Foreign exchange reserves, 2024 Q4:
  - US dollar 57.8; Euro 19.8; Japanese yen 5.8; British pound 4.7; Chinese RMB 2.2
- Outstanding international debt securities, 2024 Q4:
  - US dollar 65.2; Euro 22.8; Japanese yen 1.5; British pound 4.6; Chinese RMB 1
  - Note: Home-currency issuance is excluded.
- Outstanding international loans/deposits, 2023 Q4:
  - US dollar 57.3; Euro 19.2; Japanese yen 5.5; British pound 5.5; Chinese RMB --
- Outstanding portfolio assets, 2023:
  - US dollar 43.8; Euro 19.9; Japanese yen 7; British pound 7; Chinese RMB 1.7
- Global payment currency (SWIFT), 2024 m12:
  - US dollar 60.1; Euro 12.8; Japanese yen 5.1; British pound 4.9; Chinese RMB 2.8
- OTC FX transactions, 2022:
  - US dollar 88; Euro 31; Japanese yen 17; British pound 13; Chinese RMB 7
  - Note: Given each FX transaction involves two currencies, the currency shares add up to 200 percent.
- Liabilities of non-banks outside the issuer economy, 2024 Q4:
  - US dollar 11.9; Euro 4.2; Japanese yen 0.4; British pound --; Chinese RMB --

### 2 – Current State of International Trade and Finance (percent of global GDP; time indicated)
- Cross-Border Trade (Imports + Exports), 2023:
  - United States 14; Euro Area 15.3; Japan 4; United Kingdom 3.6; China 16.1
- Cross-Border Finance (Foreign Assets + Foreign Liabilities), 2023:
  - United States 17; Euro Area 14.9; Japan 3.4; United Kingdom 6.8; China 3.1
- Equity Market capitalization, 2024 Q3:
  - United States 51.3; Euro Area 8.4; Japan 5.6; United Kingdom 3.1; China 9.6
- Debt outstanding (all debt securities), 2024 Q3:
  - United States 51.9; Euro Area 22.9; Japan 10.5; United Kingdom 5.3; China 22.9

### 3 – Summary Statistics (percent unless otherwise stated)
- Asymmetry Index (sums to 1), 2023:
  - United States 0.42; Euro Area 0.07; Japan 0.045; United Kingdom 0.1; China 0.08
- Currency blocs (based on Frankel and Wei, 1994), 2024:
  - US dollar 47; Euro 29; Japanese yen 9; British pound 15; Chinese RMB --
- Composite Index of Currency Usage, 2023:
  - US dollar 52.7; Euro 21.3; Japanese yen 7; British pound 6.3; Chinese RMB 3.6
- Global Balance Sheet index (sums to 1), 2023:
  - US dollar 0.45; Euro 0.18; Japanese yen 0.05; British pound 0.06; Chinese RMB 0.03

---

### 5. Data Sources and Methodology (selected details)

### 1. Portfolio investment
- Source: Portfolio Investment Positions by Counterpart Economy dataset (formerly CPIS) of the IMF; includes bilateral cross-border holdings of portfolio investment securities.
- Mirror data: Used to fill missing asset holdings by using destination-country liabilities when source-country asset reports are missing; where both asset and liability data exist, the maximum value is selected (Brei and von Peter 2018).
- Intra-EMU holdings are excluded by aggregating individual EMU countries on the issuer side.
- Portfolio holdings adjusted for OFCs using a rescaling matrix from Coppola, Maggiori, Neiman and Schreger (2021) to transform residency basis to nationality basis.
  - Adjustment matrices available from 2007 to 2020; constant restatement factors used to extend sample backward and forward (following Miranda-Agrippino, Nenova, and Rey (2025)).
- Currency composition: Derived from CPIS currency dimension; when CPIS currency data is unavailable, supplemented with estimates from Allen and Juvenal (2025) extended with Milesi-Ferretti (2025).
- For the euro area currency composition: first sum individual EMU countries’ currency compositions; then exclude intra-EMU holdings using estimates of bilateral composition from Coppola and others (2021). Constant restatement factors used to extend coverage.

### 2. Foreign Direct Investment (FDI)
- Source: IMF’s Direct Investment Positions by Counterpart Economy (formerly CDIS).
- Mirror data used to maximize coverage (Broner and others 2023).
- Intra-euro area holdings are excluded and individual EMU economies aggregated.
- OFC adjustment: Identifies ultimate ownership links and excludes Special Purpose Entities (SPEs) using OECD data; employs firm-level data from Orbis to determine immediate and ultimate investors (methodology following Tan (2024) and Damgaard and others (2024)).

### 3. International banks’ loans and deposits
- Source: Locational Banking Statistics (LBS) of the BIS; records outstanding international banks’ assets and liabilities on a residency basis, including currency composition.
- Mirror data used to fill gaps as in portfolio investment.
- Intra-euro cross-border positions excluded using publicly available and confidential BIS data; EMU aggregate computed by summing individual EMU country holdings.
- No OFC adjustments due to limited nationality-basis coverage in BIS LBS (2012-onwards).
- Currency shares derived from reporting authorities’ breakdowns; minimum reported currencies include domestic currency, CHF, EUR, GBP, JPY, USD, and other foreign currencies.
- For CNY, currency data available only from 2015 onwards; currency shares of all financial instruments used as proxy for loans and deposits.

### 4. International debt securities
- Source: International Debt Securities (IDS) of the BIS; compiles outstanding amounts of bonds issued outside the local market with a complete currency breakdown.
- A debt security is considered international if the residence of the immediate issuer differs from the location of registration, governing law, or listing location.
- Home-currency issuance is excluded.

---

### 5. Trade

### Bilateral trade data
- Merchandise bilateral trade is sourced from Gaulier, G. and Zignago, S. (2010), the “BACI database”.
- BACI provides detailed information on bilateral trade flows over time on a yearly basis at the product level.
- Trade flows have been aggregated at the country-country level, and intra-EMU flows are excluded.

### Currency composition and trade invoicing
- Trade invoicing data is based on estimated invoice shares of imports and exports from Boz et al (2022), updated and expanded in Boz et al (2025).
- Data categories: USD, EUR, home currency, and other currencies.
- For Japan, the home currency share is allocated to the JPY share.
- For the UK, data is supplemented with information reported by His Majesty's (HM) Revenue & Customs.
- For China:
  - Because China does not participate in Boz et al (2025) data collection, the share of CNY used as settlement currency in current account in China is employed as a proxy for trade invoicing share.
  - It is assumed that the remaining share is in USD.
- Data for EU countries is adjusted based on the currency compositions of extra-EU trade reported by Eurostat.
- Missing values are interpolated or held constant at the nearest available value.
- To compute international shares, invoicing shares are scaled by trade flows from the BACI dataset.
- Information on cross border CNY settlement is sourced from the PBoC (accessed via Haver Analytics).

### Figure 2.3: Shares in global economy and currency composition in 2023 (methodology notes)
- Panel 1 represents the issuing jurisdiction’s share in:
  - global gross domestic product (GDP),
  - world trade (trade defined as the sum of exports and import (M+X), excluding intra-euro values), based on 2023 BACI data,
  - and world finance (total of financial assets plus total financial liabilities).
- Finance and GDP data are based on Milesi-Ferretti (2025) for 2023.
- Panel 2 components and sources:
  - International bank claims: BIS LBS by residence, end-2023, only loans and deposits considered.
  - Intl bonds = international debt securities outstanding, home-currency issuance excluded, amounts outstanding at end-2023.
  - FX reserves from IMF COFER: official reserves holdings at end-2023, available for eight major currencies and all other currencies combined; international shares are percentages of each currency in allocated reserves.
  - FX turnover: average daily turnover in all FX instruments (spot and FX derivatives) in April 2022, from the 2022 Triennial OTC Survey (net-net basis), divided by 2.
  - SWIFT data source: Cerutti, Firat and Hengge (2025).
- See section 5 for details on currency composition.

### Other indicators and data sources (variables used in chapter)
- Credit to non-bank borrowers
  - Source: Global Liquidity Indicators (GLI), BIS.
  - Notes: GLI tracks credit to non-bank borrowers, covering both loans extended by banks and funding from global bond markets through the issuance of international debt securities. It captures credit denominated in three major reserve currencies: the US dollar, euro and Japanese yen.
- SWIFT data
  - Source: Cerutti, Firat and Hengge (2025), SWIFT.
- Global exchange rate regimes
  - Source: Ilzetzki et al (2019).
- FX turnover
  - Source: BIS Triennial OTC survey (net-net basis).
- Digital currency market
  - Source: CoinGecko.
- Gold prices and reserves
  - Sources: International Financial Statistics (IFS); London Bullion Market Association (LBMA).

### Network analysis: undirected networks and centrality (2001–23)
- Undirected networks include both exports and imports between a pair of countries for the trade network, and the sum of assets and liabilities on the financial network.
- Financial network represents the sum of cross-border holding of portfolio assets and liabilities and banks' loans and deposit claims and liabilities.
- Mirror data are used to fill missing data whenever available.
- Portfolio holdings are corrected for the role of offshore financial centers using the rescaling matrix of Coppola and others (2021).
- Intra European Monetary Union holdings and trade are excluded.
- Figures referenced:
  - Figure 2.6.1: Country centrality in trade and financial network over 2001–23 (each dot is an annual value of the centrality measure).
  - Figure 2.6.2: Country centrality and currency use in trade and financial network in 2001 → 2023 (percent). Notes indicate measures by eigenvector centrality; EA = euro area, US = United States. Currency codes: USD, EUR, JPY, GBP, CNY.
  - Figure 2.6.3: Asymmetry indexes (see Asymmetry Index Algebra section below for definitions).

### Asymmetry Index: definitions and algebra
- Notation:
  - Let g_{i,t}^T and g_{i,t}^F denote centrality of country i in trade (T) or financial (F) network at time t.
  - Let c_{i,t}^T and c_{i,t}^F denote the share of country i’s currency in trade (T) or finance (F) at time t.
  - i ∈ {US, euro area, UK, China, Japan, and RoW}.
  - All centrality measures and currency shares are bounded between 0 and 1 and sum to 1 across the six economies.
- Pair-wise asymmetry index (distance from 45-degree line on 2-dimensional plane):
  - Asymmetry between centrality and currency share in trade or finance:
    - Define X_{t}^{T,F} = (x_{i,t}^{T,F}) = (g_{i,t}^{T,F}, c_{i,t}^{T,F}).
    - For each country/currency, let c̄_{i,t}^{T,F} = [ ( Σ_{i=1}^6 | x_{i,t}^{T,F} − x̄_{t}^{T,F} |^2 ) / 6 ]^{1/2}, which measures the distance from the 45-degree line (upper bar denotes the average over i’s).
    - The asymmetry index is Y_{t}^{T,F} = Σ_{i=1}^6 ω_{i,t}^{T,F} c̄_{i,t}^{T,F}, where ω_{i,t}^{T,F} denotes economy i’s share in global trade (T) or finance (F).
  - Asymmetry between centrality or currency share in trade and finance:
    - Define X_{t}^{b} = (x_{i,t}^{b}) = (g_{i,t}^{T}, g_{i,t}^{F}) and X_{t}^{c} = (x_{i,t}^{c}) = (c_{i,t}^{T}, c_{i,t}^{F}).
    - For each country/currency, let c̄_{i,t}^{c,b} = [ ( Σ_{i=1}^6 | x_{i,t}^{c,b} − x̄_{t}^{c,b} |^2 ) / 6 ]^{1/2}, which measures the distance from the 45-degree line.
    - The asymmetry index is Y_{t}^{c,b} = Σ_{i=1}^6 ω_{i,t} c̄_{i,t}^{c,b}, where ω_{i,t} = 1/2 (s_{i,t}^T + s_{i,t}^F), with s_{i,t}^T and s_{i,t}^F denoting economy i’s share in global trade and finance, respectively.
- Overall asymmetry index (distance from main diagonal in 4-dimensional space):
  - Let X_{t} = (x_{i,t}) = (g_{i,t}^T, g_{i,t}^F, c_{i,t}^T, c_{i,t}^F) and x̄_{t} = 1/4 (g_{i,t}^T + g_{i,t}^F + c_{i,t}^T + c_{i,t}^F).
  - For each country/currency, let c̄_{i,t} = [ ( Σ_{i=1}^6 | x_{i,t} − x̄_{t} |^2 /4 ) / 6 ]^{1/2}, which measures the distance from the main diagonal (in 4-dimensions).
  - The overall asymmetry index is Y_{t} = Σ_{i=1}^6 ω_{i,t} c̄_{i,t}.

### Keywords referenced (Figure 2.1 list)
- International monetary state, dominance of the US dollar, BRICs currency, Bretton Woods, Gold standard, Triffin Dilemma, Global monetary order, International financial system, Plaza Accord, International Economic cooperation, Louvre Accord, Convertibility of the dollar, Dollar parity, Reserve assets, World monetary system, Gold parity, Devaluation of the dollar, Dollar devaluation, Devalue the dollar, Dollar crisis, Currency realignment, International Monetary Crisis, Global dollar, Global role of the dollar, Trade imbalances, US dollar role, Global trading system, Exorbitant privilege, Kindleberger, US exceptionalism, Dollar status, Dollar system, Global currency, Mar-a-Lago Accord, Dollar reserve status, Central bank reserves

*External Sector Report Chapter 2 — Online Annexes (figures, tables, and methodological notes as provided).*

### 1.    Currencies of Global Balance Sheets

### 1. Currencies of Global Balance Sheets

### Method for Global Balance Sheet (GBS) index
- GBS index for currency i is defined as (퐴퐴_i + 퐿퐿_i) / (퐴퐴 + 퐿퐿), where 퐴퐴_i = foreign assets in currency i, 퐿퐿_i = foreign liabilities in currency i, and 퐴퐴 + 퐿퐿 = total foreign assets and liabilities across all currencies.
- For both assets and liabilities, estimated currency shares are used for each asset class (FDI, banking, debt, and reserves) multiplied by the share of that asset class in total foreign assets and liabilities.
- Currency weights by asset class are held constant at 2020 values (latest observation based on Allen and Juvenal, 2025) and multiplied by holdings using the latest vintage of foreign assets and liabilities data (Milesi-Ferretti, 2025 and extended with balance of payments data).
- For the euro area, intra-EU holdings are excluded.
- Currency weights are based on a confidential survey combined with publicly available data (see Allen and Juvenal (2025)).

### Outcome
- Figure 2.1.1 (not reproduced here) shows the resulting GBS index constructed as described.

---

### 2. Frankel and Wei Based Global Currency Blocs

### Methodology (two main steps)
- Step 1: Rolling monthly regressions of log returns of each currency s on log returns of global currencies (US dollar, euro, Japanese yen, British pound) expressed in the SDR numeraire:
  - Equation estimated: ∆log(s_GSSD)_t = α + Σ β_s t i ∆log(global_currency_i_GSSD)_t + ε_t (notation as in source).
  - The renminbi is excluded from estimation because it has been stable vis-à-vis the dollar in most sample years.
  - Following Ito and McCauley (2019) and Vicquéry (2022), all negative estimated coefficients are set to zero.
- Step 2: World divided into GDP-weighted currency zones according to annual co-movement of each economy’s currency with global currencies; economies can be apportioned to multiple global currencies in a given year.
  - Reserve-issuing bloc is assigned to its own currency bloc.
  - Currency weights for each economy are multiplied by GDP in US dollars and aggregated to the global economy.

### Outcome
- Figure 2.1.2 (not reproduced here) shows the resulting currency blocs over time.
- Note: Methodology builds on Frankel and Wei (1994) and follows practices in Ito and McCauley (2019), Vicquéry (2022), and related literature.

---

### 3. Network Centrality Measures

### Eigenvector centrality used to assess importance in trade and financial networks
- Centrality score c_i for country i defined by: λ c_i = Σ g_j i c_j for all i ∈ N (notation as in source), where g denotes the adjacency matrix capturing connection strength.
- In directed networks, g = g_ji represents exports from country j to i or j’s holdings of i’s assets, expressed as a share of world GDP.
- In undirected networks, g = exports + imports between i and j, or sum of cross-holdings of financial assets and liabilities.
- λ is the largest eigenvector; centrality scores are normalized to sum to 1 (Σ c_j = 1).
- Interpretation: A country’s centrality is proportional to the sum of the centrality of its neighbors; linkages with more important countries receive higher weight.

---

### 4. Snapshot of the International Monetary System (selected statistics)

### 1 – Current State of International Currencies (percent of total; time indicated)
- Foreign exchange reserves, 2024 Q4:
  - US dollar 57.8; Euro 19.8; Japanese yen 5.8; British pound 4.7; Chinese RMB 2.2
- Outstanding international debt securities, 2024 Q4:
  - US dollar 65.2; Euro 22.8; Japanese yen 1.5; British pound 4.6; Chinese RMB 1
  - Note: Home-currency issuance is excluded.
- Outstanding international loans/deposits, 2023 Q4:
  - US dollar 57.3; Euro 19.2; Japanese yen 5.5; British pound 5.5; Chinese RMB --
- Outstanding portfolio assets, 2023:
  - US dollar 43.8; Euro 19.9; Japanese yen 7; British pound 7; Chinese RMB 1.7
- Global payment currency (SWIFT), 2024 m12:
  - US dollar 60.1; Euro 12.8; Japanese yen 5.1; British pound 4.9; Chinese RMB 2.8
- OTC FX transactions, 2022:
  - US dollar 88; Euro 31; Japanese yen 17; British pound 13; Chinese RMB 7
  - Note: Given each FX transaction involves two currencies, the currency shares add up to 200 percent.
- Liabilities of non-banks outside the issuer economy, 2024 Q4:
  - US dollar 11.9; Euro 4.2; Japanese yen 0.4; British pound --; Chinese RMB --

### 2 – Current State of International Trade and Finance (percent of global GDP; time indicated)
- Cross-Border Trade (Imports + Exports), 2023:
  - United States 14; Euro Area 15.3; Japan 4; United Kingdom 3.6; China 16.1
- Cross-Border Finance (Foreign Assets + Foreign Liabilities), 2023:
  - United States 17; Euro Area 14.9; Japan 3.4; United Kingdom 6.8; China 3.1
- Equity Market capitalization, 2024 Q3:
  - United States 51.3; Euro Area 8.4; Japan 5.6; United Kingdom 3.1; China 9.6
- Debt outstanding (all debt securities), 2024 Q3:
  - United States 51.9; Euro Area 22.9; Japan 10.5; United Kingdom 5.3; China 22.9

### 3 – Summary Statistics (percent unless otherwise stated)
- Asymmetry Index (sums to 1), 2023:
  - United States 0.42; Euro Area 0.07; Japan 0.045; United Kingdom 0.1; China 0.08
- Currency blocs (based on Frankel and Wei, 1994), 2024:
  - US dollar 47; Euro 29; Japanese yen 9; British pound 15; Chinese RMB --
- Composite Index of Currency Usage, 2023:
  - US dollar 52.7; Euro 21.3; Japanese yen 7; British pound 6.3; Chinese RMB 3.6
- Global Balance Sheet index (sums to 1), 2023:
  - US dollar 0.45; Euro 0.18; Japanese yen 0.05; British pound 0.06; Chinese RMB 0.03

---

### 5. Data Sources and Methodology (selected details)

### 1. Portfolio investment
- Source: Portfolio Investment Positions by Counterpart Economy dataset (formerly CPIS) of the IMF; includes bilateral cross-border holdings of portfolio investment securities.
- Mirror data: Used to fill missing asset holdings by using destination-country liabilities when source-country asset reports are missing; where both asset and liability data exist, the maximum value is selected (Brei and von Peter 2018).
- Intra-EMU holdings are excluded by aggregating individual EMU countries on the issuer side.
- Portfolio holdings adjusted for OFCs using a rescaling matrix from Coppola, Maggiori, Neiman and Schreger (2021) to transform residency basis to nationality basis.
  - Adjustment matrices available from 2007 to 2020; constant restatement factors used to extend sample backward and forward (following Miranda-Agrippino, Nenova, and Rey (2025)).
- Currency composition: Derived from CPIS currency dimension; when CPIS currency data is unavailable, supplemented with estimates from Allen and Juvenal (2025) extended with Milesi-Ferretti (2025).
- For the euro area currency composition: first sum individual EMU countries’ currency compositions; then exclude intra-EMU holdings using estimates of bilateral composition from Coppola and others (2021). Constant restatement factors used to extend coverage.

### 2. Foreign Direct Investment (FDI)
- Source: IMF’s Direct Investment Positions by Counterpart Economy (formerly CDIS).
- Mirror data used to maximize coverage (Broner and others 2023).
- Intra-euro area holdings are excluded and individual EMU economies aggregated.
- OFC adjustment: Identifies ultimate ownership links and excludes Special Purpose Entities (SPEs) using OECD data; employs firm-level data from Orbis to determine immediate and ultimate investors (methodology following Tan (2024) and Damgaard and others (2024)).

### 3. International banks’ loans and deposits
- Source: Locational Banking Statistics (LBS) of the BIS; records outstanding international banks’ assets and liabilities on a residency basis, including currency composition.
- Mirror data used to fill gaps as in portfolio investment.
- Intra-euro cross-border positions excluded using publicly available and confidential BIS data; EMU aggregate computed by summing individual EMU country holdings.
- No OFC adjustments due to limited nationality-basis coverage in BIS LBS (2012-onwards).
- Currency shares derived from reporting authorities’ breakdowns; minimum reported currencies include domestic currency, CHF, EUR, GBP, JPY, USD, and other foreign currencies.
- For CNY, currency data available only from 2015 onwards; currency shares of all financial instruments used as proxy for loans and deposits.

### 4. International debt securities
- Source: International Debt Securities (IDS) of the BIS; compiles outstanding amounts of bonds issued outside the local market with a complete currency breakdown.
- A debt security is considered international if the residence of the immediate issuer differs from the location of registration, governing law, or listing location.
- Home-currency issuance is excluded.

*External Sector Report Chapter 2 — Online Annexes (figures, tables, and methodological notes as provided).*

### 5. Trade

### 5. Trade

### Bilateral trade data
- Merchandise bilateral trade is sourced from Gaulier, G. and Zignago, S. (2010), the “BACI database”.
- BACI provides detailed information on bilateral trade flows over time on a yearly basis at the product level.
- Trade flows have been aggregated at the country-country level, and intra-EMU flows are excluded.

### Currency composition and trade invoicing
- Trade invoicing data is based on estimated invoice shares of imports and exports from Boz et al (2022), updated and expanded in Boz et al (2025).
- Data categories: USD, EUR, home currency, and other currencies.
- For Japan, the home currency share is allocated to the JPY share.
- For the UK, data is supplemented with information reported by His Majesty's (HM) Revenue & Customs.
- For China:
  - Because China does not participate in Boz et al (2025) data collection, the share of CNY used as settlement currency in current account in China is employed as a proxy for trade invoicing share.
  - It is assumed that the remaining share is in USD.
- Data for EU countries is adjusted based on the currency compositions of extra-EU trade reported by Eurostat.
- Missing values are interpolated or held constant at the nearest available value.
- To compute international shares, invoicing shares are scaled by trade flows from the BACI dataset.
- Information on cross border CNY settlement is sourced from the PBoC (accessed via Haver Analytics).

### Figure 2.3: Shares in global economy and currency composition in 2023 (methodology notes)
- Panel 1 represents the issuing jurisdiction’s share in:
  - global gross domestic product (GDP),
  - world trade (trade defined as the sum of exports and import (M+X), excluding intra-euro values), based on 2023 BACI data,
  - and world finance (total of financial assets plus total financial liabilities).
- Finance and GDP data are based on Milesi-Ferretti (2025) for 2023.
- Panel 2 components and sources:
  - International bank claims: BIS LBS by residence, end-2023, only loans and deposits considered.
  - Intl bonds = international debt securities outstanding, home-currency issuance excluded, amounts outstanding at end-2023.
  - FX reserves from IMF COFER: official reserves holdings at end-2023, available for eight major currencies and all other currencies combined; international shares are percentages of each currency in allocated reserves.
  - FX turnover: average daily turnover in all FX instruments (spot and FX derivatives) in April 2022, from the 2022 Triennial OTC Survey (net-net basis), divided by 2.
  - SWIFT data source: Cerutti, Firat and Hengge (2025).
- See section 5 for details on currency composition.

### Other indicators and data sources (variables used in chapter)
- Credit to non-bank borrowers
  - Source: Global Liquidity Indicators (GLI), BIS.
  - Notes: GLI tracks credit to non-bank borrowers, covering both loans extended by banks and funding from global bond markets through the issuance of international debt securities. It captures credit denominated in three major reserve currencies: the US dollar, euro and Japanese yen.
- SWIFT data
  - Source: Cerutti, Firat and Hengge (2025), SWIFT.
- Global exchange rate regimes
  - Source: Ilzetzki et al (2019).
- FX turnover
  - Source: BIS Triennial OTC survey (net-net basis).
- Digital currency market
  - Source: CoinGecko.
- Gold prices and reserves
  - Sources: International Financial Statistics (IFS); London Bullion Market Association (LBMA).

### Network analysis: undirected networks and centrality (2001–23)
- Undirected networks include both exports and imports between a pair of countries for the trade network, and the sum of assets and liabilities on the financial network.
- Financial network represents the sum of cross-border holding of portfolio assets and liabilities and banks' loans and deposit claims and liabilities.
- Mirror data are used to fill missing data whenever available.
- Portfolio holdings are corrected for the role of offshore financial centers using the rescaling matrix of Coppola and others (2021).
- Intra European Monetary Union holdings and trade are excluded.
- Figures referenced:
  - Figure 2.6.1: Country centrality in trade and financial network over 2001–23 (each dot is an annual value of the centrality measure).
  - Figure 2.6.2: Country centrality and currency use in trade and financial network in 2001 → 2023 (percent). Notes indicate measures by eigenvector centrality; EA = euro area, US = United States. Currency codes: USD, EUR, JPY, GBP, CNY.
  - Figure 2.6.3: Asymmetry indexes (see Asymmetry Index Algebra section below for definitions).

### Asymmetry Index: definitions and algebra
- Notation:
  - Let g_{i,t}^T and g_{i,t}^F denote centrality of country i in trade (T) or financial (F) network at time t.
  - Let c_{i,t}^T and c_{i,t}^F denote the share of country i’s currency in trade (T) or finance (F) at time t.
  - i ∈ {US, euro area, UK, China, Japan, and RoW}.
  - All centrality measures and currency shares are bounded between 0 and 1 and sum to 1 across the six economies.
- Pair-wise asymmetry index (distance from 45-degree line on 2-dimensional plane):
  - Asymmetry between centrality and currency share in trade or finance:
    - Define X_{t}^{T,F} = (x_{i,t}^{T,F}) = (g_{i,t}^{T,F}, c_{i,t}^{T,F}).
    - For each country/currency, let c̄_{i,t}^{T,F} = [ ( Σ_{i=1}^6 | x_{i,t}^{T,F} − x̄_{t}^{T,F} |^2 ) / 6 ]^{1/2}, which measures the distance from the 45-degree line (upper bar denotes the average over i’s).
    - The asymmetry index is Y_{t}^{T,F} = Σ_{i=1}^6 ω_{i,t}^{T,F} c̄_{i,t}^{T,F}, where ω_{i,t}^{T,F} denotes economy i’s share in global trade (T) or finance (F).
  - Asymmetry between centrality or currency share in trade and finance:
    - Define X_{t}^{b} = (x_{i,t}^{b}) = (g_{i,t}^{T}, g_{i,t}^{F}) and X_{t}^{c} = (x_{i,t}^{c}) = (c_{i,t}^{T}, c_{i,t}^{F}).
    - For each country/currency, let c̄_{i,t}^{c,b} = [ ( Σ_{i=1}^6 | x_{i,t}^{c,b} − x̄_{t}^{c,b} |^2 ) / 6 ]^{1/2}, which measures the distance from the 45-degree line.
    - The asymmetry index is Y_{t}^{c,b} = Σ_{i=1}^6 ω_{i,t} c̄_{i,t}^{c,b}, where ω_{i,t} = 1/2 (s_{i,t}^T + s_{i,t}^F), with s_{i,t}^T and s_{i,t}^F denoting economy i’s share in global trade and finance, respectively.
- Overall asymmetry index (distance from main diagonal in 4-dimensional space):
  - Let X_{t} = (x_{i,t}) = (g_{i,t}^T, g_{i,t}^F, c_{i,t}^T, c_{i,t}^F) and x̄_{t} = 1/4 (g_{i,t}^T + g_{i,t}^F + c_{i,t}^T + c_{i,t}^F).
  - For each country/currency, let c̄_{i,t} = [ ( Σ_{i=1}^6 | x_{i,t} − x̄_{t} |^2 /4 ) / 6 ]^{1/2}, which measures the distance from the main diagonal (in 4-dimensions).
  - The overall asymmetry index is Y_{t} = Σ_{i=1}^6 ω_{i,t} c̄_{i,t}.

### Keywords referenced (Figure 2.1 list)
- International monetary state, dominance of the US dollar, BRICs currency, Bretton Woods, Gold standard, Triffin Dilemma, Global monetary order, International financial system, Plaza Accord, International Economic cooperation, Louvre Accord, Convertibility of the dollar, Dollar parity, Reserve assets, World monetary system, Gold parity, Devaluation of the dollar, Dollar devaluation, Devalue the dollar, Dollar crisis, Currency realignment, International Monetary Crisis, Global dollar, Global role of the dollar, Trade imbalances, US dollar role, Global trading system, Exorbitant privilege, Kindleberger, US exceptionalism, Dollar status, Dollar system, Global currency, Mar-a-Lago Accord, Dollar reserve status, Central bank reserves

*External Sector Report Chapter 2 — IMF | July 2025; content unit: 5. Trade (ch2annex - 5. Trade).*

---


_Source: https://www.imf.org/-/media/files/publications/esr/2025/english/ch2annex.pdf_