## wpiea2020249-print-pdf

## Source details

**Canonical URL:** [wpiea2020249-print-pdf](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020249-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020249-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020249-print-pdf.pdf.json)

---

### Data sample and scope
- China: "the second largest economy in the world" and "the largest banking system in the world" with "$35 trillion in total assets (about 300 percent of China’s GDP)".
- Four Chinese banks among the 30 largest G-SIBs (Financial Stability Board, November 2019).
- Chinese banks represent "about 7 percent of total cross-border bank lending" and reported claims on "196 out of 216 borrower countries" (BIS LBS, as of mid-2018).
- Data coverage: bilateral relationships between up to "39 lender countries" and up to "185 borrower countries" (31 AEs, 143 EMDE, and 11 offshore centres). China became a BIS reporting country in 2016 (data starting as of end-2015).

### Nationality perspective, data composition, and novel measures
- Nationality (ultimate owner) perspective used; rationale:
  - Across bank nationalities, about "60 percent" of cross-border lending is extended from the home country; remainder from affiliates abroad.
  - For EMDE banks lending to other EMDEs, home-country offices grant on average "~33 percent" of total cross-border claims.
- Cross-border lending composition:
  - Loans: "about 67 percent" of total claims.
  - Debt securities: "about 22 percent" as of Q3:2019.
- Weighted distance measure (novel):
  - Dist_l b = sum_r Dist_r b * (XBC_l r b / XBC_l b), where Dist_r b is capital-to-capital distance between residence r and borrower b; XBC_l r b is cross-border lending from residence r by nationality l to borrower b; XBC_l b is total cross-border lending by nationality l to borrower b.
  - Backflows dropped from computation.
- Traditional gravity controls: geographical distance (capitals), colonial relationship (post-1945, 0/1), common language (0/1).
- Bilateral lagged economic ties: trade (sum of exports and imports), outward FDI (ultimate investor perspective), total portfolio investment (debt + equity from IMF CPIS). Lagged values (end-2016) used to mitigate endogeneity.

### Major empirical findings — distance, complementarities, and distributional facts
- Distance effects:
  - "1 percent rise in bilateral borrower-lender distance is associated with a 0.62 percent decline in cross-border lending" (all borrowers and lenders).
  - AE vs EMDE heterogeneity (baseline cross-sectional results):
    - AE borrowers: "1 percent" increase in weighted distance → "0.43 percent" drop in outstanding cross-border claims (Table 4, column 2).
    - EMDE borrowers: "1 percent" increase in weighted distance → "0.67 percent" drop (Table 4, column 4).
  - Weighted distance generally increases the estimated importance of distance relative to simple capital-to-capital distance; weighted-distance coefficient often more negative and statistically significant (sometimes at the "10 percent" level for AE borrowers).
- Trade, portfolio investment, and FDI as complements:
  - Trade:
    - AE borrowers: "1 percent" rise in bilateral trade → "0.77 percent" increase in cross-border lending (baseline).
    - EMDE borrowers: "1 percent" rise in trade → "0.63 percent" increase.
  - Portfolio investment:
    - "1 percent" increase → "0.35 percent" rise in lending to AE borrowers; "0.31 percent" rise for EMDE borrowers (baseline).
  - FDI:
    - Weak impact: around "0.05−0.06 percent" and only marginally significant for EMDE borrowers (baseline).
  - Common colony and common language indicators add little once distance, trade, portfolio investment and FDI are controlled for.
- Distributional statistics (bilateral pairs, all currencies):
  - Average outstanding stock: "reaches US$4 billion"; median "US$20 million".
  - Bilateral trade: average "about US$5 billion"; median "US$243 million".
  - Bilateral portfolio investments: average "about US$8 billion"; median "US$12 million".
  - "14 percent" of bilateral pairs share a common language; "2 percent" linked by historical colonial ties.

### Findings specific to Chinese banks and comparisons by nationality
- Market shares and creditor status:
  - Chinese banks are the most important creditor for "66 (out of 185)" borrower countries (overall sample).
  - Individual-country market shares (largest reported): Japanese "15.4 percent", U.S. "11.3 percent", French "11.3 percent", U.K. "9.5 percent", Chinese "7.1 percent", Swiss "6.4 percent".
  - For EMDE borrowers, Chinese banks extend about "24 percent" of total cross-border bank lending (highest share); Japanese banks second at "11.2 percent".
  - Out of "143 EMDEs", Chinese banks are the most important creditor for "63" borrower countries.
- Chinese banks’ behavior:
  - Expansion resembles global reach of AE banks when lending to EMDEs; resemblance stronger for U.S. dollar-denominated claims.
  - Chinese banks perceive distance to EMDE borrowers as less of a barrier than other EMDE banks; act more like U.S. and European banks on distance sensitivity.
  - Trade-lending: Chinese banks show a particularly strong positive correlation between bilateral trade and cross-border lending to EMDEs — stronger than Japanese and European banks and close to U.S. banks.
  - Portfolio-lending: unlike other banking systems, past portfolio investment is negatively correlated with cross-border lending to EMDE borrowers for Chinese banks (attributed to capital outflow restrictions and concentrated portfolio investment in a few AEs). When lending to AE borrowers, Chinese banks show strong complementarities with portfolio investment.
  - FDI-lending: only weak evidence of a relationship for Chinese banks.
- Swap lines and BRI:
  - Presence of swap lines with the People’s Bank of China may help explain some variation in Chinese banks’ lending; Belt and Road Initiative membership is not statistically significant once trade is controlled for.
- Country- and currency-specific magnitudes:
  - Chinese and U.S. banks: sum of standalone and interaction coefficients implies a 1 percent increase in cross-border trade is almost matched by a 1 percent increase in cross-border lending to EMDEs (Chinese and U.S. banks).
  - For Chinese banks lending to AEs: a 1 percent rise in bilateral trade raises outstanding claims by "0.4 percent".
  - For U.S. dollar claims to EMDEs, the net effect of a 1 percent increase in bilateral distance for Chinese banks is a "0.94 percent" decline in U.S. dollar claims.

### Regression evidence (selected exact estimates)
- Table 3 (2018 cross-section, 39 lending parent countries, up to 185 borrowers) — Simple Distance specification:
  - Simple Distance coefficients (All counterparties): -1.673*** (0.134); AEs: -0.854*** (0.130); EMDEs: -1.306*** (0.131).
  - Trade (All counterparties): 0.731*** (0.055); AEs: 0.638*** (0.058); EMDEs: 0.670*** (0.060).
  - Investment (All counterparties): 0.365*** (0.044).
  - FDI (All counterparties): 0.079*** (0.029).
  - Adjusted R-square: 0.708 (column 1).
- Table 4 (specification (2), by lending nationality) — Interaction Trade*CN(0/1), Investment*CN(0/1), FDI*CN(0/1):
  - Trade*CN(0/1) — Borrowers EMDEs: 0.452*** (0.083) (simple distance).
  - Investment*CN(0/1) — Borrowers EMDEs: -0.601*** (0.062) (simple distance).
  - FDI*CN(0/1) — Borrowers EMDEs: -0.170*** (0.053) (simple distance).
  - Observations: Advanced 1,071; EMDEs 3,700. Adjusted R-squared: Advanced 0.840; EMDEs 0.688.
- Table 5 (U.S. dollar lending, specification (2)):
  - Trade standalone (USD lending): AEs: 0.416** (0.159); EMDEs: 0.293*** (0.089).
  - Interaction Trade*CN(0/1) (USD lending): EMDEs: 0.876*** (0.111).
  - Interaction FDI*CN(0/1) (USD lending): EMDEs: -0.248*** (0.053).
  - Observations: Advanced 942; EMDEs 2,481. Adjusted R-squared: Advanced 0.713; EMDEs 0.583.
- Table 6 (Chinese banks only, EMDE borrowers, 2018 cross-section):
  - Trade: consistently ~0.92−0.95 with significance (e.g., 0.935*** (0.190)).
  - Distance standalone estimates vary and are imprecise (examples: -0.524 (0.472); -0.337 (0.665)).
  - Sample: Observations 137; R-squared range: 0.354–0.373.
- Table 7 (PPML and Keep Zeros robustness):
  - Distance (PPML / Keep Zeros selected):
    - Col4 EMDEs weighted: -0.352** (0.170).
    - Col8 EMDEs weighted ln(Y+1): -0.205* (0.115).
  - Trade (selected): Col1 0.574*** (0.125); Col4 0.275*** (0.103); Col8 0.271*** (0.046).
  - Investment (selected): Col1 0.343*** (0.083); Col4 0.367*** (0.058); Col8 0.329*** (0.051).
  - FDI (selected): Col3 0.109** (0.053); Col4 0.104** (0.049); Col8 0.167*** (0.053).
  - Dist*CN(0/1) interaction examples: Col1 -0.089 (0.255); Col4 0.005 (0.175); Col6 0.178** (0.085); Col8 -0.303** (0.116).
  - R-squared (PPML): AEs 0.958 and 0.956; EMDEs 0.819 and 0.832.

### Robustness, zeros treatment, and methodology notes
- Zeros: Many zero observations due to panel construction (185 borrowers × 39 lender nationalities). Missing BIS values treated as zeros.
- Two approaches to treat zeros:
  - PPML estimator (Santos Silva and Tenreyro, 2006) using outstanding values including zeros (Table 7, cols 1–4).
  - Keep zeros by using ln(Y+1) as dependent variable (Table 7, cols 5–8).
- Both approaches confirm main results: negative distance coefficients and positive correlations for trade, portfolio investment, and FDI, mostly statistically significant—especially for EMDE borrowers.
- Fixed effects: All columns include separate lending parent and borrower country fixed effects (LC+BC).
- Standard errors clustered by lending parent country; significance: *** p<0.01, ** p<0.05, * p<0.1.

### Policy implications, scenarios, and research directions
- Trade shocks:
  - A reduction in global trade (e.g., shortening of value chains due to trade tensions and/or COVID-19) could be associated with a decline in Chinese cross-border bank lending, especially to EMDEs given strong trade-lending correlation.
- Capital account liberalization:
  - Further integration of Chinese equity and bond markets could move Chinese banks’ correlation between cross-border lending and portfolio investment closer to other nationalities and may increase their cross-border lending reach.
- Ownership and crisis behavior:
  - Government ownership shares in large Chinese banks may affect geopolitical and economic behavior (e.g., state-owned banks lending more during crises), warranting further research.
- Data and methodological limits:
  - Short BIS reporting history for China (reporting from Q4:2015) constrains time-series analysis; longer time series would improve inference about dynamics and policy channels.

### Numerical highlights (preserve exact reported values)
- "1 percent" rise in bilateral borrower-lender distance → "0.62 percent" decline in cross-border lending (all borrowers and lenders).
- AE borrowers: "1 percent" increase in weighted distance → "0.43 percent" drop (Table 4, column 2).
- EMDE borrowers: "1 percent" increase in weighted distance → "0.67 percent" drop (Table 4, column 4).
- AE trade elasticity: "0.77 percent" (1 percent rise in trade → 0.77 percent rise in lending).
- EMDE trade elasticity: "0.63 percent".
- Portfolio investment elasticity: "0.35 percent" (AEs); "0.31 percent" (EMDEs).
- FDI elasticity: "0.05−0.06 percent" (weak).
- Bilateral claims averages and medians: average "reaches US$4 billion"; median "US$20 million".
- China-specific statistics:
  - Chinese banks top creditor in "66 (out of 185)" countries (overall) and in "63" EMDEs (out of 143).
  - Chinese banks account for "24 percent" of cross-border lending to EMDEs.
  - Chinese banks’ cross-border lending equals "18 percent" of China's GDP.
  - Chinese banks direct "43 percent" of their total lending to EMDE borrowers.
- Selected regression coefficients (exact preserved examples):
  - Simple Distance (All counterparties): -1.673*** (0.134).
  - Trade (All counterparties, simple distance): 0.731*** (0.055).
  - Investment (All counterparties, simple distance): 0.365*** (0.044).
  - FDI (All counterparties, simple distance): 0.079*** (0.029).
  - Table 4 Interaction Investment*CN(0/1) — EMDEs: -0.601*** (0.062).
  - Table 5 Interaction Trade*CN(0/1) — EMDEs (USD lending): 0.876*** (0.111).
  - Chinese banks distance effect on USD claims to EMDEs: "0.94 percent" decline per 1 percent increase in bilateral distance.

*Source: wpiea2020249-print-pdf (IMF Working Paper content provided in the prompt).*

### References  ............................................................................................................

### wpiea2020249-print-pdf - References  ...............................................................................................................

### Figures
- 1. Simple and Weighted Distance Measures ..................................................................... 23
- 2A. Market Shares in Different Types of International Economic Interactions with AEs ..... 24
- 2B. Market Shares in Different Types of International Economic Interactions  
  with EMDEs .......................................................................................................... 25

### Tables
- 1. Measures of Global Relevance by Bank Nationality: Top Cross-Border Creditors  
  and Market Share................................................................................................... 26
- 2. Descriptive Statistics .................................................................................................... 27
- 3. Cross-Border Lending, Distance and Other Types of Economic Interaction .................. 28
- 4. Do Results Differ by Nationality of the Lending Bank? ................................................ 29
- 5. Do Results Differ by Nationality of the Lending Bank? U.S. Dollar Lending ................ 31
- 6. Zooming in on Chinese Banks ...................................................................................... 33
- 7. Alternative Estimation Methods: PPML and Keeping Zeros ......................................... 34

### Annex Tables
- 1. Cross-Border Banking, Trade and Investment: China Compared with Other Major  
  Players in Advanced Economies (AEs) and EMDEs .............................................. 36
- 2. Variable Description ..................................................................................................... 41

*Source: wpiea2020249-print-pdf - References  ....................................................................................................................... 44*

### 3. Data Sample .........................................................................................................

### 3. Data Sample

### I. Context and scope
- China is described as "the second largest economy in the world" and "the largest banking system in the world" with "$35 trillion in total assets (about 300 percent of China’s GDP)".  
- Four Chinese banks feature among the 30 largest global systemically important banks (G-SIBs) as communicated by the Financial Stability Board after consultations with the BCBS and national authorities in November 2019.  
- As of mid-2018, Chinese banks represent "about 7 percent of total cross-border bank lending" and reported claims on "196 out of 216 borrower countries" (BIS LBS).  
- Data coverage for the analysis: bilateral relationships between up to "39 lender countries" and up to "185 borrower countries" (31 AEs, 143 EMDE, and 11 offshore centres). China became a BIS reporting country in 2016 (with data starting as of end-2015).

### II. Nationality perspective and data composition
- The analysis uses the nationality (ultimate owner) perspective of BIS international banking statistics rather than bank residence; rationale:
  - Across all bank nationalities, about "60 percent" of their cross-border lending is extended from their home country, with the remainder from affiliates abroad.  
  - For EMDE banks lending to other EMDEs, "on average, offices in the home country grant only about one third of their total cross-border claims", i.e., ~33 percent.  
- Cross-border lending definition and composition:
  - Includes bank loans, holdings of debt securities and other instruments.
  - Loans constitute "about 67 percent" of total claims.
  - Debt securities account for "about 22 percent" as of Q3:2019.
- The residence perspective can distort bilateral lender-borrower relationships; e.g., Chinese affiliates in Hong Kong SAR may be classified differently in some BIS aggregates.

### III. Empirical approach (gravity framework and weighted distance)
- Methodology:
  - Cross-sectional gravity models with lender- and borrower-country fixed effects to absorb lender- and borrower-specific characteristics.
  - Bilateral lagged economic ties included: trade (sum of exports and imports), outward FDI (adjusted to ultimate investor perspective), and total portfolio investment (debt plus equity from IMF CPIS).
  - Lagged values of trade, portfolio investment and FDI are used to mitigate endogeneity concerns.
- Traditional gravity controls: geographical distance (between capitals), colonial relationship (post-1945, 0/1), common language (0/1).
- Novel weighted distance measure:
  - Accounts for the full network of affiliates: for a parent country l and borrower b, Dist_l b = sum over residences r of Dist_r b * (XBC_l r b / XBC_l b), where XBC_l r b is cross-border lending from residence r by nationality l to borrower b and XBC_l b is total cross-border lending by nationality l to borrower b.
  - Bilateral distance Dist_r b captures distances between capitals of borrower b and residence r.
  - Backflows (claims on the home country intermediated by offices abroad) are dropped from the computation.

### IV. Key empirical findings — information asymmetries and complementarities
- Distance effects:
  - Across all borrowers and lenders, a "1 percent rise in bilateral borrower-lender distance is associated with a 0.62 percent decline in cross-border lending."  
  - Distance deters lending to EMDE borrowers relatively more than to AE borrowers. Example estimates:
    - For AE borrowers, a 1 percent increase in weighted distance is associated with a "0.43 percent" drop in outstanding cross-border claims (Table 4, column 2).
    - For EMDE borrowers, a 1 percent increase in weighted distance is associated with a "0.67 percent" drop (Table 4, column 4).
  - The weighted distance measure typically increases the estimated importance of distance relative to the simple bilateral capital-to-capital distance; the weighted measure captures affiliate networks and often yields stronger negative distance-lending correlations. For AE borrowers the weighted-distance coefficient turns negative and significant at the "10 percent" level in some specifications.
- Trade, portfolio investment, and FDI as complements:
  - Trade and other bilateral economic ties are positively correlated with cross-border lending, interpreted as complementarities that can reduce information asymmetries.
  - Magnitudes (from baseline cross-sectional results):
    - For AE borrowers, a "1 percent" rise in bilateral trade is associated with a "0.77 percent" increase in cross-border lending.
    - For EMDE borrowers, a "1 percent" rise in trade is associated with a "0.63 percent" increase.
    - A "1 percent" increase in total portfolio investment is associated with a "0.35 percent" rise in lending to AE borrowers, and a "0.31 percent" rise for EMDE borrowers.
    - The impact of FDI is weaker: around "0.05−0.06 percent" and only marginally significant for EMDE borrowers.
  - Traditional common colony and common language indicators add little explanatory power once distance, trade, portfolio investment and FDI are controlled for.
- Distributional statistics and skew:
  - For bilateral cross-border claims (all currencies): average outstanding stock "reaches US$4 billion", median is "US$20 million".
  - Bilateral trade: average "about US$5 billion", median "US$243 million".
  - Bilateral portfolio investments: average "about US$8 billion", median "US$12 million".
  - "14 percent" of bilateral pairs share a common language; "2 percent" are linked by historical colonial ties.

### V. Findings specific to Chinese banks and comparisons across nationalities
- Market shares and creditor status:
  - Chinese banks are the most important creditor for "66 (out of 185)" borrower countries — more than any other banking system in the sample.
  - Individual-country market shares (largest reported): Japanese "15.4 percent", U.S. "11.3 percent", French "11.3 percent", U.K. "9.5 percent", Chinese "7.1 percent", Swiss "6.4 percent".
  - For EMDE borrowers, Chinese banks extend about "24 percent" of total cross-border bank lending (the highest share). Japanese banks are second at "11.2 percent".
  - Out of "143 EMDEs", Chinese banks are the most important creditor for "63" borrower countries.
- Chinese banks’ behavior:
  - Chinese banks’ expansion resembles the global reach of banks from AEs when lending to EMDEs; this resemblance is more pronounced for claims denominated in U.S. dollars.
  - Chinese banks appear to perceive distance to EMDE borrowers as less of a barrier than other EMDE banks; they act more like U.S. and European banks in this respect.
  - Trade-lending relationship: Chinese banks show a particularly strong positive correlation between bilateral trade and cross-border lending to EMDEs — stronger than Japanese and European banks and closer to U.S. banks.
  - Portfolio-lending relationship: unlike all other banking systems, past portfolio investment is negatively correlated with cross-border lending to EMDE borrowers for Chinese banks — attributed to China’s capital outflow restrictions and concentrated portfolio investment within a few AEs. When lending to AE borrowers, Chinese banks show strong complementarities with portfolio investment.
  - FDI-lending relationship: only weak evidence of a relationship for Chinese banks.
- Drivers and policy-relevant interactions:
  - The presence of swap lines between the People’s Bank of China and a borrower country (but not Belt and Road Initiative membership) might help explain some variation in Chinese banks’ lending patterns.
  - Potential macro scenarios:
    - A prospective reduction in global trade (e.g., shortening of value chains due to trade tensions and/or COVID-19) could be associated with a decline in Chinese cross-border bank lending, especially to EMDEs, given the strong trade-lending correlation.
    - Further integration of Chinese equity and bond markets could move Chinese banks’ correlation between cross-border lending and portfolio investment closer to other bank nationalities and may increase their cross-border lending reach.

### VI. Methodological contributions and robustness
- Contributions:
  - Empirical demonstration of the importance of a nationality (ultimate owner) perspective for analyzing global cross-border banking, especially given multinational affiliate networks and offshore financial centres.
  - Introduction and use of a weighted distance measure that accounts for lenders’ affiliate locations; the weighted measure more often yields the expected negative sign for distance in lending regressions than the simple capital-to-capital distance.
  - Joint testing of trade, FDI, and portfolio investment as complements to cross-border bank lending within a gravity framework.
- Robustness and checks:
  - Results are robust to addressing sample bias from zero bilateral lending using a balanced panel and PPML estimator (Santos Silva and Tenreyro, 2006) — see robustness checks referenced in the paper.
  - Main findings hold when isolating U.S. dollar claims and when varying parent-country groupings; China is always kept separate in robustness checks.

*Source: IMF Working Paper (section "3. Data Sample" and surrounding text from the provided PDF content).*

### 0.57 percent rise in lending to AEs(Table 4, column 2), and a 0.52 percent raise in lending to

### wpiea2020249-print-pdf - 0.57 percent rise in lending to AEs(Table 4, column 2), and a 0.52 percent raise in lending to

### Key findings on cross-border bank lending and economic ties
- A 0.57 percent rise in lending to AEs (Table 4, column 2), and a 0.52 percent raise in lending to EMDEs (Table 4, column 4), respectively.
- Trade and portfolio investment generally complement bilateral cross-border bank lending relationships.
- Distance, used as a proxy for information asymmetries, negatively affects cross-border lending, with heterogeneous effects by parent bank nationality and borrower type (AE vs EMDE).

### Bank-nationality specific relationships
- Chinese and U.S. banks: the sum of standalone and interaction coefficients implies that a 1 percent increase of cross-border trade is almost matched with a 1 percent increase in cross-border lending to EMDEs.
- Chinese banks vs U.S. banks: Chinese banks’ sensitivity of cross-border lending to EMDEs is similar to U.S. banks with regard to trade, despite most Chinese banks being state-owned.
- European and Japanese lenders: complementarity between trade and lending is weaker than for Chinese and U.S. banks.
- Other EMDE banks: show the weakest complementarity between trade and lending.
- For lending to AE borrowers, effects are less pronounced for Chinese, U.S., and Japanese banks; specifically, for Chinese banks a 1 percent rise in bilateral trade raises their outstanding claims by only 0.4 percent.

### Role of portfolio investment
- On average, a 1 percent higher level of portfolio investment is associated with a 0.37−0.38 percent higher level of outstanding cross-border claims for both AEs and EMDEs (standalone coefficient).
- China deviates: for EMDE counterparties, bilateral claims extended by Chinese banks are 0.21 percent lower when portfolio investment increases.
- For AE counterparties, a 1 percent rise in bilateral portfolio investment is associated with Chinese banks reporting a 0.76 percent higher level of outstanding bilateral amounts.
- Interpretation: lower market share and concentration of Chinese portfolio investment on AE countries, and capital outflow controls in China, likely drive contrasts; information asymmetries may not be reduced for Chinese banks through portfolio investment.

### FDI relationships with cross-border bank lending
- On average, a 1 percent increase in FDI is associated with a 0.11 percent rise in cross-border bank lending, independent of whether the borrower resides in an AE or an EMDE (Table 4).
- Heterogeneity for EMDE counterparties:
  - Japanese banks: a 1 percent rise in bilateral FDI is associated with a 0.17 percent decline in lending to EMDE borrowers originated by Japanese banks.
  - Chinese banks: the decline is about 0.05 percent.
  - U.S. banks: the decline is about 0.03 percent.
- For AE borrowers, the net effect of bilateral FDI is marginally positive for Chinese, U.S., and other EMDE banks, but almost zero for European and Japanese banks.

### U.S. dollar denominated cross-border bank lending (subset analysis)
- Restricting to U.S. dollar denominated claims generally weakens most relationships slightly, but strengthens core findings for Chinese banks.
- Distance:
  - For AE borrowers, the effect of distance turns insignificant on average for U.S. dollar claims.
  - For EMDE borrowers, distance remains only marginally significant.
- Trade vs portfolio investment by borrower type:
  - For U.S. dollar lending to AEs, trade continues to play the most important role in offsetting information asymmetries.
  - For lending to EMDE borrowers in U.S. dollars, the relationship between portfolio investment becomes more important.
- Chinese banks and distance (U.S. dollar claims):
  - The net effect of a 1 percent increase in bilateral distance to an EMDE borrower is associated with a 0.94 percent decline in U.S. dollar claims for Chinese banks.
  - This magnitude is similar to the net effect exhibited by banks from other EMDEs.
  - Japanese banks: bilateral distance has an even more negative impact on their lending to EMDEs.
- For borrowers in AE, there is no effect of distance on U.S. dollar lending for Chinese or European banks.
- Japanese banks: in U.S. dollar claims, the net effect of distance on lending to AEs is positive, explained by the geographical structure of their affiliate network.
- U.S. banks: U.S. dollar-denominated claims are special because U.S. banks act as key funding sources in their home currency.

### Numerical highlights (preserve exact reported values)
- 0.57 percent (rise in lending to AEs, Table 4, column 2).
- 0.52 percent (raise in lending to EMDEs, Table 4, column 4).
- 1 percent increase of cross-border trade ≈ 1 percent increase in cross-border lending (Chinese and U.S. banks to EMDEs).
- 0.4 percent (Chinese banks: increase in outstanding claims for a 1 percent rise in bilateral trade to AE borrowers).
- 0.37−0.38 percent (average association of 1 percent higher portfolio investment with higher outstanding cross-border claims).
- 0.21 percent lower (Chinese banks’ bilateral claims to EMDE counterparties when portfolio investment increases).
- 0.76 percent higher (Chinese banks’ outstanding bilateral amounts for a 1 percent rise in portfolio investment to AE counterparties).
- 0.11 percent (average association of a 1 percent increase in FDI with cross-border bank lending).
- 0.17 percent decline (Japanese banks: effect of a 1 percent rise in bilateral FDI on lending to EMDE borrowers).
- 0.05 percent decline (Chinese banks: effect of a 1 percent rise in bilateral FDI on lending to EMDE borrowers).
- 0.03 percent decline (U.S. banks: effect of a 1 percent rise in bilateral FDI on lending to EMDE borrowers).
- 0.94 percent decline (Chinese banks: net effect of a 1 percent increase in bilateral distance to an EMDE borrower on U.S. dollar claims).

*Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020249-print-pdf.pdf*

### 0.6 percent in the case of U.S. banks and Japanese banks. This difference between Chinese

### wpiea2020249-print-pdf

### Key empirical findings on cross-border bank lending and economic ties
- Chinese banks’ cross-border lending reaches 18 percent of the country’s GDP.
- Chinese banks direct 43 percent of their total lending to EMDE borrowers.
- Chinese banks account for 24 percent of all cross-border lending to EMDE borrowers; Japanese banks account for 11 percent of the EMDE total.
- Chinese banks are the top creditor in 63 borrower countries (out of 135 EMDE borrower countries in the sample).
- Across all bank nationalities, lenders are more sensitive to distance when lending to EMDEs than when lending to AEs; negative coefficients for distance persist after controlling for other bilateral ties.
- Trade, portfolio investment, and FDI frequently act as complements to cross-border lending, with past trade in particular showing a positive correlation with cross-border lending.

### Differences by currency denomination and nationality
- The positive complementarities between trade and cross-border lending are stronger for Chinese banks when lending to EMDEs; this difference is even larger for U.S. dollar claims than for total cross-border claims.
- For lending to borrowers in AEs, complementarities with portfolio investment are particularly strong for Chinese banks.
- Evidence that Chinese FDI complements cross-border banking with AEs is present, but this effect disappears when isolating cross-border claims denominated in U.S. dollars.

### Zooming in on China-specific policies and their effects
- Analysis restricted to Chinese banks and EMDE borrowers uses dummies for participation in the BRI (as of 2015) and for bilateral swap-line arrangements with the PBOC.
- In the Chinese-only EMDE subset, trade is the only economic-relationship variable with a positive and statistically significant association with cross-border bank lending.
- The BRI indicator is not statistically significant either standalone or in interaction; the swap-line indicator is only marginally significant.
- The lack of significance for BRI may stem from its positive correlation with trade, with trade absorbing potential BRI effects.

### Robustness checks and treatment of zeros
- The dataset construction produces many zeros because 185 potential borrower countries are allocated to each of 39 bank nationalities; missing values in the BIS input indicate no outstanding amounts (treated as zeros).
- To address potential bias from zeros in log-linear gravity models, two approaches were used:
  - Poisson Pseudo-Maximum Likelihood (PPML) estimator (Table 7, columns 1 to 4).
  - Adding a small positive dollar amount to zero outstanding relationships to retain those observations in log specifications (Table 7, columns 5 to 8).
- Both robustness approaches confirm main results: negative coefficients for distance and positive correlation coefficients for trade, portfolio investment, and FDI, mostly statistically significant—especially for EMDE borrowers.
- Specific results for Chinese banks remain similar under robustness checks: Chinese banks, relative to other nationalities, lend relatively more to EMDE borrowers with higher bilateral trade; distance is more negatively correlated with Chinese bank lending to EMDE borrowers.

### Data scope, novel measures, and methodological contributions
- Analysis uses BIS locational banking statistics by bank nationality, combined with CEPII distance data and data on trade and international capital flows.
- Unique features:
  - Nationality perspective capturing cross-border operations of home offices and global affiliate networks.
  - Creation of a weighted distance measure that sums across all locations from which a given bank nationality extends claims to a borrower country, weighting by the relative importance of each location in the lender-borrower relationship (formula presented in text).
  - Dataset allows distinguishing more than 5,000 lender-borrower pairs across AEs and EMDEs.
- Descriptive statistics sample sizes and selected moments (all counterparties):
  - Observations: 5,090
  - Total lending, ln(X): N 5,090, mean 2.69, p50 2.97, sd 4.32, min -6.91, max 14.00
  - USD lending, ln(X): N 3,712, mean 3.01, p50 3.40, sd 3.92, min -6.91, max 13.93
  - Trade, ln(X+1): mean 5.48, p50 5.50, sd 2.70, min 0.00, max 13.30
  - Investment, ln(X+1): mean 3.39, p50 2.53, sd 3.56, min 0.00, max 14.30
  - FDI, ln(X+1): mean 1.84, p50 0.00, sd 3.14, min 0.00, max 14.00
  - Simple distance, ln(X): mean 8.59, p50 8.84, sd 0.85, min 4.09, max 9.89
  - Weighted distance, all, ln(X): mean 8.52, p50 8.74, sd 0.85, min 0.00, max 9.89

### Regression evidence (selected results preserved exactly)
- Table 3 (specification (1), 2018 cross-section, 39 lending parent countries, up to 185 borrowers)
  - Simple Distance coefficients (All counterparties): -1.673*** (0.134); for AEs: -0.854*** (0.130); for EMDEs: -1.306*** (0.131).
  - Trade coefficient (All counterparties, simple distance specification): 0.731*** (0.055); for AEs: 0.638*** (0.058); for EMDEs: 0.670*** (0.060).
  - Investment coefficient (All counterparties, simple distance): 0.365*** (0.044).
  - FDI coefficient (All counterparties, simple distance): 0.079*** (0.029).
  - Adjusted R-square reported: 0.708 (column 1).
- Table 4 (specification (2), by lending nationality)
  - Standalone Trade coefficients:
    - Borrowers Advanced: 0.597*** (0.123) using Simple distance.
    - Borrowers EMDEs: 0.504*** (0.081) using Simple distance.
  - Interaction Trade*CN(0/1):
    - Borrowers Advanced: -0.097 (0.078) simple distance;
    - Borrowers EMDEs: 0.452*** (0.083) simple distance.
  - Interaction Investment*CN(0/1):
    - Borrowers Advanced: 0.312*** (0.099) simple distance;
    - Borrowers EMDEs: -0.601*** (0.062) simple distance.
  - Interaction FDI*CN(0/1):
    - Borrowers Advanced: 0.121 (0.085) simple distance;
    - Borrowers EMDEs: -0.170*** (0.053) simple distance.
  - Observations: Advanced 1,071; EMDEs 3,700. Adjusted R-squared: Advanced 0.840 (col 1), EMDEs 0.688 (col 3).
- Table 5 (U.S. dollar lending, specification (2))
  - Standalone Trade coefficients (USD lending):
    - Borrowers Advanced: 0.416** (0.159) simple distance.
    - Borrowers EMDEs: 0.293*** (0.089) simple distance.
  - Interaction Trade*CN(0/1) (USD lending):
    - Borrowers Advanced: -0.251** (0.114) simple distance.
    - Borrowers EMDEs: 0.876*** (0.111) simple distance.
  - Interaction FDI*CN(0/1) (USD lending):
    - Borrowers EMDEs: -0.248*** (0.053) simple distance.
  - Observations: Advanced 942; EMDEs 2,481. Adjusted R-squared: Advanced 0.713 (col 1), EMDEs 0.583 (col 3).

### Policy implications and potential channels
- If trade tensions or persistent declines in global trade (for example, from COVID-19 aftermath) occur, cross-border bank lending could fall in parallel; the decline in global banking could be more pronounced for Chinese banks because they display significantly higher correlations with trade than some peers.
- Liberalization reforms in the Chinese bond market could foster further inward and outward portfolio investment; if portfolio liberalization makes China more similar to other AE and EMDE countries, Chinese banks’ investments abroad could increase to diversify, potentially lowering information asymmetries for Chinese cross-border bank lending.
- Ownership structure in China (government’s considerable ownership shares in largest Chinese banks) could affect geopolitical and economic behavior of banks (e.g., state-owned banks lending more during crises to stabilize the economy), suggesting avenues for future research.

### Limitations and suggestions for future research
- Cross-sectional empirical approach constrained by a short reporting history: China started reporting to BIS locational banking statistics with positions from Q4:2015, limiting time-series analysis.
- Future research could benefit from a longer time series and further exploration of China-specific factors such as state ownership and geopolitical objectives.

*Source: wpiea2020249-print-pdf (IMF Working Paper content provided in the prompt).*

### Section II and Graph 1. All columns include separate lending parent and borrower country fixed effects. EU captures the 

### Section II and Graph 1. All columns include separate lending parent and borrower country fixed effects. EU captures the

### Estimation notes and methodology
- All columns include separate lending parent and borrower country fixed effects (LC+BC).
- Simple Distance and Weighted Distance are described in Section II and Graph 1.
- R2_all-FF shows the difference between the adjusted R2 based on the full specification and a regression with only CP and BC fixed effects; reported R2_all-FF values: 0.0719, 0.0729, 0.152, 0.152 (as listed in Table 7).
- Standard errors in parentheses are clustered by lending parent country (Table 6: robust standard errors); significance: *** p<0.01, ** p<0.05, * p<0.1.
- Samples:
  - Table 6: 2018 cross-section of Chinese lending parent banks and all EMDE borrower countries (BCs); Observations: 137; R-squared range: 0.354–0.373.
  - Table 7: 2018 cross-section with 39 parent countries (LCs); Observations: PPML columns 1+2: 1,187; PPML columns 3+4: 5,565; Keep Zeros columns 5+6: 1,187; Keep Zeros columns 7+8: 5,565.
  - Table 7 borrower counts: BCs 31 (AEs) and 143 (EMDEs); LCs 39.

### Table 6 — Zooming in on Chinese Banks (Dependent Variable: Total lending in all currencies (ln), all control variables)
- Sample restricted to Chinese parent banks (2018 cross-section) and EMDE borrower countries.
- Observations: 137 in each column; R-squared: columns 1–4 around 0.354–0.356; columns 5–8 around 0.368–0.373.
- Standalone variable coefficients (selected exact estimates and standard errors):
  - Distance: -0.524 (0.472), -0.337 (0.665), -0.600 (0.498), -0.432 (0.755), -0.200 (0.381), -0.210 (0.593), -0.280 (0.382), -0.345 (0.547).
  - X (BRI 2015 or Swap line indicator): -0.302 (0.649), 3.590 (10.688), -0.363 (0.674), 2.585 (10.532), 1.359* (0.720), 0.437 (7.583), 1.360* (0.715), -0.942 (7.419).
  - Trade: 0.935*** (0.190), 0.922*** (0.248), 0.936*** (0.189), 0.920*** (0.248), 0.932*** (0.190), 0.952*** (0.208), 0.928*** (0.189), 0.946*** (0.207).
  - Investment: -0.038 (0.185), -0.037 (0.338), -0.042 (0.184), -0.034 (0.336), -0.150 (0.206), -0.217 (0.237), -0.149 (0.203), -0.213 (0.231).
  - FDI: 0.032 (0.133), 0.047 (0.223), 0.031 (0.131), 0.043 (0.219), 0.042 (0.134), 0.027 (0.157), 0.037 (0.132), 0.016 (0.154).
- Interaction effects (selected exact estimates and standard errors):
  - X*Distance: -0.460 (1.060), -0.357 (1.057), 0.103 (0.693), 0.249 (0.680).
  - X*Trade: 0.032 (0.441), 0.043 (0.433), -0.163 (0.556), -0.156 (0.556).
  - X*FDI: -0.018 (0.293), -0.021 (0.287), 0.002 (0.322), 0.015 (0.320).
  - X*Investment: -0.011 (0.391), -0.025 (0.388), 0.370 (0.434), 0.365 (0.431).
- Note: Except for the indicators X, all explanatory variables enter the analysis in log. Annex Table 2 gives precise definitions; Table A3 lists the countries grouped into EMDEs and AEs.

### Table 7 — Alternative Estimation Methods: PPML and Keeping Zeros (Dependent Variable: Total lending in all currencies (ln) to all sectors)
- Two estimation strategies:
  - PPML estimator using outstanding values including zeros as dependent variable (Columns 1–4).
  - Keeping zeros by using ln(Y+1) as dependent variable (Columns 5–8).
- Columns alternate between simple distance (cols 1+2, 5+6) and weighted distance (cols 3+4, 7+8).
- Fixed Effects: LC+BC in all columns.
- Selected standalone variable estimates (exact values and standard errors):
  - Distance (PPML / Keep Zeros):
    - Col1 AEs simple: 0.130 (0.159)
    - Col2 EMDEs simple: -0.017 (0.215)
    - Col3 AEs weighted: -0.087 (0.207)
    - Col4 EMDEs weighted: -0.352** (0.170)
    - Col5 AEs simple ln(Y+1): -0.365 (0.242)
    - Col6 EMDEs simple ln(Y+1): -0.212 (0.181)
    - Col7 AEs weighted ln(Y+1): -0.152 (0.134)
    - Col8 EMDEs weighted ln(Y+1): -0.205* (0.115)
  - Trade:
    - Col1: 0.574*** (0.125); Col2: 0.468*** (0.118); Col3: 0.304*** (0.101); Col4: 0.275*** (0.103); Col5: 0.298** (0.112); Col6: 0.312*** (0.115); Col7: 0.273*** (0.047); Col8: 0.271*** (0.046).
  - Investment:
    - Col1: 0.343*** (0.083); Col2: 0.340*** (0.087); Col3: 0.367*** (0.054); Col4: 0.367*** (0.058); Col5: 0.341*** (0.084); Col6: 0.331*** (0.080); Col7: 0.325*** (0.050); Col8: 0.329*** (0.051).
  - FDI:
    - Col1: 0.034 (0.039); Col2: 0.036 (0.041); Col3: 0.109** (0.053); Col4: 0.104** (0.049); Col5: 0.167** (0.070); Col6: 0.174** (0.068); Col7: 0.171*** (0.052); Col8: 0.167*** (0.053).
- Selected interaction coefficients (exact values and standard errors):
  - Dist*CN(0/1): -0.089 (0.255), 0.347 (0.286), -0.216 (0.218), 0.005 (0.175), 0.993*** (0.274), 0.178** (0.085), -0.322** (0.133), -0.303** (0.116).
  - Dist*US(0/1): 0.157 (0.164), -0.002 (0.184), 0.163 (0.335), 0.151 (0.193), 0.714*** (0.261), -0.060 (0.073), -0.063 (0.173), -0.203*** (0.055).
  - Dist*JP(0/1): 1.003* (0.565), -0.332 (0.455), 0.210 (0.208), 0.267* (0.150), -2.359*** (0.728), 1.908*** (0.183), 0.055 (0.139), -0.678*** (0.068).
  - Trade*CN(0/1): -0.125 (0.088), -0.131 (0.089), 0.166* (0.096), 0.169* (0.099), 0.017 (0.075), 0.009 (0.079), 0.551*** (0.045), 0.553*** (0.044).
  - Investment*CN(0/1): 0.109 (0.073), 0.021 (0.062), -0.135*** (0.044), -0.143*** (0.053), 0.209*** (0.076), 0.240*** (0.071), -0.340*** (0.051), -0.342*** (0.051).
  - FDI*CN(0/1): -0.053 (0.079), -0.001 (0.065), -0.222*** (0.064), -0.215*** (0.061), 0.026 (0.077), -0.006 (0.073), -0.166*** (0.047), -0.161*** (0.049).
  - Investment*JP(0/1): 0.295*** (0.088), 0.398*** (0.103), 0.212*** (0.061), 0.224*** (0.068), 0.531*** (0.072), 0.422*** (0.074), 0.172*** (0.046), 0.199*** (0.049).
  - FDI*JP(0/1): -0.023 (0.052), -0.021 (0.058), -0.131** (0.061), -0.160*** (0.062), -0.160** (0.067), -0.237*** (0.067), -0.161*** (0.051), -0.244*** (0.058).
- Goodness of fit:
  - R-squared reported in Table 7: PPML AEs: 0.958 and 0.956; PPML EMDEs: 0.819 and 0.832.
- Notes:
  - Columns 1+3: borrowers in up to 31 advanced economies; columns 2+4: up to 143 EMDEs.
  - Except for parent bank indicators (e.g., CN(0/1) for China), all bilateral explanatory variables are in logs.
  - Standard errors clustered by lender parent country.

### Annex Table 1 — Cross-Border Banking, Trade and Investments: China Compared with Other Major Players in AEs and EMDEs
- Scope: Comparative positions for multiple countries across metrics: Total bank lending (2018Q2), USD-denominated bank lending (2018Q2), Imports (End 2016), Exports (End 2016), FDI (End 2015), Portfolio investment (End 2017); sources: BIS locational banking statistics (by nationality); CEPII database; authors’ calculations.
- Selected exact figures (China and comparators; shares and counts preserved as in table):
  - Total bank lending (2018Q2) — Share in global lending:
    - Among AE borrowers: CN 2.4; JP 14.8; FR 13.6; GB 10.4; US 11.1; DE 10.3; ES 3.0; CH 5.8.
    - Among EMDE borrowers: CN 23.7; JP 11.2; FR 8.1; GB 7.8; US 7.1; DE 5.1; ES 3.1; CH 3.1.
    - Among OFC borrowers: CN 13.7; JP 24.3; FR 5.6; GB 4.9; US 13.4; DE 6.0; ES 0.8; CH 11.4.
    - Top lender (# countries) among AE borrowers: CN 0; JP 3; FR 6; GB 0; US 1; DE 4; ES 2; CH 1.
  - USD-denominated bank lending (2018Q2) — Among EMDE borrowers: CN 63; JP 6; FR 10; GB 5; US 9; DE 5; ES 2; CH 4.
    - Share in global lending among EMDE borrowers: CN 27.2; JP 10.6; FR 6.4; GB 7.8; US 9.2; DE 3.5; ES 2.8; CH 2.9.
    - Top lender (# countries) among EMDE borrowers: CN 55; JP 9; FR 8; GB 6; US 14; DE 4; ES 1; CH 6.
  - Imports (End 2016) — Share in global imports among EMDE borrowers: CN 12.9; US 21.7; JP 6.5; FR 2.7; GB 2.9; DE 6.8; ES 1.9; CH 1.3.
    - Top partner (# countries) among EMDE borrowers: CN 41; US 32; JP 5; FR 5; GB 1; DE 8; ES 3; CH 6.
  - Exports (End 2016) — Share in global exports among EMDE borrowers: CN 16.3; US 13.9; JP 6.8; FR 2.5; GB 2.0; DE 8.8; ES 1.4; CH 1.6.
    - Top partner (# countries) among EMDE borrowers: CN 56; US 28; JP 1; FR 8; GB 0; DE 11; ES 2; CH 0.
  - FDI (End 2015) — Share in global FDI among EMDE borrowers: CN 14.1; US 12.2; JP 0.4; FR 0.9; GB 1.5; DE 1.9; ES 1.9; CH 0.1.
    - Top lender (# countries) among EMDE borrowers: CN 16; US 5; JP 0; FR 4; GB 4; DE 2; ES 1; CH 0.
  - Portfolio investment (End 2017) — Share in global investment among EMDE borrowers: CN 0.4; US 35.4; JP 3.6; FR 1.8; GB 6.5; DE 3.1; ES 0.4; CH 1.5.
    - Top lender (# countries) among EMDE borrowers: CN 0; US 62; JP 1; FR 5; GB 5; DE 1; ES 0; CH 2.
- Additional country panels in Annex Table 1 provide analogous breakdowns for NL, IT, CA, AT, AU, BE, SE, DK, PT, GR, CY, NO, LU, FI, IE, KR, BR, IN, RU, ZA, TR, MY, PH, MX, ID, CL, SG, HK, PA, BH, MO, TT and smallest-economy panels for OFC and certain financial centers; values are preserved exactly in the table (see Annex Table 1).

*Source: wpiea2020249-print-pdf - Section II and Graph 1. All columns include separate lending parent and borrower country fixed effects. EU captures the*

### Annex Table 2. Variable Description

### Annex Table 2. Variable Description

### Variable definitions
- Total lending  
  - Cross-border lending, all currencies  
  - Unit/Value: In of USD mn  
  - Source/Comment: BIS locational banking statistics (by nationality)

- USD lending  
  - Cross-border lending denominated in USD  
  - Unit/Value: In of USD mn  
  - Source/Comment: BIS locational banking statistics (by nationality)

- Colony  
  - Indicator of colonial relationship between pairs of countries (after 1945)  
  - Unit/Value: 0/1  
  - Source/Comment: CEPII database

- Language  
  - Indicator of Common Official Language (using at least 9% of the population between two countries)  
  - Unit/Value: 0/1  
  - Source/Comment: CEPII database

- Trade  
  - Sum of exports and imports  
  - Unit/Value: ln of USD mn  
  - Source/Comment: UN Comtrade Database

- Investment  
  - Total Portfolio Investment (Equity+Debt)  
  - Unit/Value: ln of USD mn  
  - Source/Comment: IMF, CPIS

- FDI  
  - Foreign direct investment  
  - Unit/Value: ln of USD mn  
  - Source/Comment: Damgaard and Elkjaer (2017)

- SDistance  
  - Simple distance between capitals of country pairs  
  - Unit/Value: ln of km  
  - Source/Comment: CEPII database

- WDistance  
  - Weighted distance between capitals of country pairs  
  - Unit/Value: ln of km  
  - Source/Comment: With weights as share of lending from different locations of bank’s affiliates at home and abroad. See Graph 3 for a more detailed description.

### Notes on data treatment and coverage
- The aggregation by nationality in the case of Chinese banks captures only cross-border claims of those banks from Mainland China and their global network of affiliates. It does not include, for example, Chinese banks, which are domestically owned in Hong Kong SAR.
- We use end-2016 data as lagged values for the variables on bilateral economic ties (Trade, Investment and FDI) to mitigate endogeneity concerns.
- In order to retain observations with small values of bilateral ties, we use ln(X+1) in the regressions.

*Source: Annex Table 2. Variable Description (from provided PDF content)*

### Annex Table 3. Data Sample — key listings and counts
- Parent Countries/Jurisdictions: 39, with the Number of Counterparty Borrower Countries shown in parentheses for each:
  - AU: Australia (70)  
  - AT: Austria(103)  
  - BE: Belgium(135)  
  - BH: Bahrain  
  - BR: Brazil(64)  
  - CA: Canada(51)  
  - CH: Switzerland(16)  
  - CL: Chile(38)  
  - CN: China(148)  
  - CY: Cyprus(93)  
  - DE: Germany(113)  
  - DK: Denmark(51)  
  - ES: Spain(14)  
  - FI: Finland(16)  
  - FR: France(56)  
  - GB: United Kingdom(80)  
  - GR: Greece(84)  
  - HK: Hong Kong SAR(14)  
  - ID: Indonesia(54)  
  - IE: Ireland(59)  
  - IN: India(148)  
  - IT: Italy(143)  
  - JP: Japan(30)  
  - KR: Korea(140)  
  - LU: Luxembourg(30)  
  - MX: Mexico(23)  
  - MO: Macao SAR(18)  
  - MY: Malaysia (15)  
  - NL: Netherlands(39)  
  - NO: Norway(73)  
  - PA: Panama(38)  
  - PH: Philippines(50)  
  - PT: Portugal(20)  
  - RU: Russia(81)  
  - SE: Sweden(90)  
  - SG: Singapore(49)  
  - TR: Turkey(101)  
  - US: United States(113)  
  - ZA: South Africa(87)

- Cross-Border Counterparty Borrower Countries/Jurisdictions: 185 total, split into:
  - A. Advanced Economies (AEs, 31 of 36 jurisdictions) — listing includes AU, AT, BE, CA, CY, CH, DK, EE, FI, FR, DE, GR, IE, IS, IT, JP, LU, LV, LT, MT, NL, NZ, NO, PT, SM, SI, SK, ES, SE, GB, US
  - B. Offshore Financial Centres (OFCs, 11 of 21 jurisdictions) — includes BS, BH, BB, CW, GI, HK, JE, KY, MO, MU, PA, SG, SX, VU, WS (as presented in the source)
  - C. Emerging Markets and Developing Economies (EMDEs, 143 of 159 jurisdictions) — extensive listing including AF, AL, AM, AO, AR, AZ, BD, BEJ?, etc., and many others as provided in the source. (The source contains the full EMDE list and country annotations.)

- Special markers in lists:
  - Note: # indicates Swap line agreement with China; a total of 50 countries.
  - * indicates BRI 2015 countries, and the list also includes Bhutan (BT), Iran (IR), Maldives (MV), Palestine Territory (PS), Syria (SY), Timor Leste (TL) and Turkmenistan (TM); a total of 65 countries.

### Excluded borrower countries/jurisdictions (counts and shares)
- List of 31 Excluded Borrower Countries/Jurisdictions (=216-185)
  - AEs: Advanced Economies (5) — AD: Andorra; FO: Faeroe Islands; GL: Greenland; LI: Liechtenstein; VA: Vatican City
  - EMDEs: Emerging markets & developing economies (16) — BQ: Bonaire, Sint Eustatius and Saba; CU: Cuba; FK: Falkland Islands; FM: Micronesia; KP: North Korea; NC: New Caledonia; PF: French Polynesia; PU: US Pacific islands; PS - Palestinian Territory; SS - South Sudan; SH - St. Helena; SZ - Swaziland; TL – Timor-Leste; TC - Turks and Caicos Islands; WF: Wallis and Futuna; and others as listed in the source.
  - Offshore centres (10) — AW: Aruba; BM: Bermuda; CW: Curacao; GI: Gibraltar; GG: Guernsey; IM: Isle of Man; JE: Jersey; JE : Jersey; KY: Cayman Islands; SX: Sint Maarten; 1Z: West Indies UK (as presented)

- Shares and aggregate figures (as stated in source)
  - Share of excluded BCs within AEs: 0.1%  
  - Share of excluded BCs within EMDEs: 1.1%  
  - Share of all 31 excluded BCs: 9.3% of total cross-border claims on all 216 BCs  
  - Share of excluded BCs within offshore centres: 56.1% (41% is only due to Cayman Islands)

*Source: Annex Table 3. Data Sample (from provided PDF content)*

*Source: Provided IMF PDF content (Annex Tables 2–3 and accompanying notes and listings).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020249-print-pdf.pdf_