## 1. Gravity Model Results on Trade and Financial Interconnectedness

## Source details

**Canonical URL:** [1. Gravity Model Results on Trade and Financial Interconnectedness](https://www.imf.org/-/media/files/publications/cr/2025/english/1sgpea2025002-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2025/english/1sgpea2025002-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2025/english/1sgpea2025002-source-pdf.pdf.json)

---

### Introduction
- Paper purpose: take stock of financial interconnectedness in Asia and through Singapore, and explore link between trade and financial flows.
- Key framing:
  - Asian economies are not as financially integrated as their integration in trade would suggest.
  - Asia’s financial integration with the rest of the world has been led by advanced economies in the region, including Singapore.
  - The FDI network closely resembles the trade network; the portfolio assets network is dominated by advanced economies.
  - A gravity-model analysis confirms that countries with strong prior trade linkages are likely to see higher FDI flows but finds no such relationship for portfolio flows.
  - Singaporean banks are exposed to cross-border and FX funding risks; authorities continue mitigation efforts.

### Asia’s Financial Integration with the Rest of the World
- High-level findings:
  - Asia and the Pacific make up over 30 percent of global trade and GDP, hold less than 30 percent of the total inward FDI stock, and make up less than 15 percent of global portfolio liabilities.
  - Asia’s share in global trade grew substantially in the late 2000s and early 2010s but has stabilized since then.
  - Asia’s share of global FDI has risen in recent years; share of global portfolio liabilities has been stable.
- Cross-border positions by country group:
  - Cross-border assets and liabilities (IIP sum) for Asian advanced economies rose from around 100 percent of GDP in 1990 to over 500 percent in 2023.
  - Emerging market economies in the region have cross-border assets and liabilities as percent of GDP that have remained relatively stable since the 1990s.
- Singapore-specific IIP patterns:
  - About half of all FDI coming to Singapore is from either Europe or the United States.
  - The United States has become increasingly important as a source of FDI inflows to Singapore; Europe’s role has diminished.
  - Outbound FDI from Singapore is mostly destined to other parts of Asia.
  - On the portfolio side, about a quarter of portfolio investments is in U.S. assets; China plays an increasing role. Portfolio liabilities are small overall.

### Network Analysis on Financial Integration in Asia
- Method:
  - Network analysis using nodes (countries) and edges (flows); eigenvector centrality used to capture influence of nodes.
- Trade network (2023 snapshot and historical evolution):
  - Rise of the Chinese mainland since the 1980s mirrored by decline of Japan.
  - India, Hong Kong SAR, Singapore, ASEAN countries present; Vietnam has risen substantially.
  - Singapore’s centrality has been on a slowly decreasing trend since peaking in the late 1990s.
- FDI (direct investment) network:
  - Resembles the trade network closely.
  - Chinese mainland attracted substantial FDI; Japan’s relative decline observed.
  - Hong Kong SAR and Singapore play larger roles than in the trade network (financial centers).
  - Singapore’s importance has grown steadily, having attracted significant FDI in recent years.
- Portfolio investment network:
  - Disconnected from trade and FDI networks.
  - Japan remains the biggest Asian holder of portfolio assets, followed by Australia.
  - India and the Chinese mainland are increasing but remain small.
  - Singapore’s centrality broadly stable.
  - Overall, portfolio assets network within Asia is dominated by advanced economies.

### Relationship between Trade and Financial Interconnectedness (Gravity Model)
- Model specification and identification:
  - Modified gravity model where financial exposures (stocks of FDI and portfolio investments) are dependent variables and trade is included on the RHS.
  - Includes origin-time (μi,t), destination-time (δj,t), and origin-destination (φi,j) fixed effects — “fully saturated” specification.
  - Trade variable is lagged by one year to address endogeneity (trade(-1)).
  - Estimation method: maximum likelihood.
  - Data sources: FDI from CDIS, portfolio investments from CPIS, trade from DOTS.
- Main empirical findings:
  - Trade is a key driver for FDI but not for portfolio investments.
  - Baseline regression results (excerpt from Table 1):
    - Trade (-1) coefficients by dependent variable:
      - FDI: 0.005*** (standard error (0.002))
      - Assets: Portfolio: -0.000 (0.001)
      - Assets: Debt: 0.001 (0.002)
      - Assets: Equity: 0.000 (0.001)
      - Liabilities: Portfolio: 0.001 (0.001)
      - Liabilities: Debt: 0.003 (0.002)
      - Liabilities: Equity: 0.002 (0.001)
    - Observations: 143,658; 156,872; 140,455; 119,888; 149,783; 135,842; 114,644 (by column).
    - Country-Year FE: Y (all columns)
    - Counterpart-Year FE: Y (all columns)
    - Country Pair FE: Y (all columns)
    - Pseudo R2: 0.984; 0.993; 0.987; 0.994; 0.993; 0.987; 0.994 (by column).
    - Standard errors clustered by country-pair.
    - Significance: * p<0.10, ** p<0.05, *** p<0.01.
  - Interpretation provided in text:
    - The coefficient for lagged trade is significant at the 99 percent level when using FDI as the dependent variable.
    - A one percentage point increase in the bilateral trade relationship (as percent of average GDP) is associated with an increase in the FDI stock by approximately 0.5 percent.
    - The coefficient for lagged trade is insignificant for portfolio investment (assets/liabilities and debt/equity).
  - Robustness:
    - Findings robust to using shares rather than stocks and to including additional lags of the trade variable.

### Cross-Border Linkages of Singaporean Banks (December 2024)
- Banking sector scale and role:
  - Singapore is an international financial center; banking sector assets totaled 512 percent of GDP in December 2024.
  - Singaporean banks function as intermediaries channeling capital between advanced economies and emerging Asian markets.
- Cross-border asset exposures:
  - Banks’ claims on non-residents totaled 116.6 percent of GDP, accounting for about 22 percent of total banking sector assets.
  - Claims composition:
    - Loans and advances including bills financing to non-residents: 67.4 percent of GDP or 12.8 percent of total banking system assets.
    - Securities and equities issued by private non-resident entities: 25.9 percent of GDP or 4.9 percent of total banking system assets.
    - Amounts due from non-resident banks: 23 percent of GDP or 4.4 percent of total banking system assets.
  - Recent change:
    - Banks’ claims on non-residents declined from 141.3 percent of GDP in December 2021 to 116.6 percent of GDP in December 2024, mostly due to decline in loans to non-residents from 89.6 percent of GDP in December 2021 to 67.4 percent of GDP in December 2024.
- Funding and liabilities:
  - Liabilities of the Singaporean banking system to non-residents totaled 194.4 percent of GDP, accounting for 36.8 percent of total banking sector liabilities.
  - Liability composition:
    - Non-bank deposits by non-residents: 86.7 percent of GDP.
    - Amounts payable to non-resident banks, including interbank loans: 107.7 percent of GDP.
- Structure of local D-SIBs (2024Q4 / December 2024):
  - Three local D-SIBs (DBS, UOB, and OCBC) accounted for approximately 52 percent of Singapore's banking sector assets in December 2024.
  - D-SIBs have regionally diversified loan portfolios and rely on deposits denominated in foreign currencies.
  - Local D-SIBs’ loans by region (2024Q4):
    - Singapore: 45 percent of total loans.
    - China (including Hong Kong SAR and Taiwan Province of China): 23 percent.
    - Asia and Pacific excluding China: 16 percent.
    - Rest of the world: 16 percent.
  - Currency composition (2024Q4):
    - About 61 percent of local D-SIBs’ loans and deposits were denominated in foreign currencies.
    - Loans denominated in U.S. dollars accounted for about 22 percent of total loans.
    - Deposits denominated in U.S. dollars accounted for about 34 percent of total deposits.
- Market context note:
  - Singapore's foreign exchange derivatives market, the third largest globally, has grown 45 percent since 2019, primarily driven by asset managers, bank treasurers, and corporate treasurers using FX swaps and cross-currency swaps (BIS Triennial Central Bank Survey, 2022).

### Exposures and Channels of Vulnerability
- Macro and funding shock exposures:
  - Banks in Singapore are exposed to global and regional macroeconomic shocks, a sharp tightening in global financial conditions entailing high global interest rates, shortage in U.S. dollar liquidity, and spillovers from banking sector stresses outside Singapore.
  - Given interconnectedness, these shocks could impact Singaporean banks and broader regional financial stability.
- Historical resilience cited:
  - Global Financial Crisis, onset of the COVID pandemic, and the banking turmoil in the U.S. and Europe in 2023.
- Distribution of exposures:
  - Loans by region (repeated): 45% Singapore; 23% China; 16% Rest of Asia-Pacific; 16% Rest of the World.

### MAS Stress Tests and Interbank Contagion Analysis
- MAS’s Industry-Wide Stress Test (November 2024):
  - Adverse scenario features intensification of geopolitical conflicts, trade tensions, and global inflationary pressures, entailing a global recession affecting countries and regions on which D-SIBs have exposures, including China and other Asian economies amid elevated interest rates.
  - In the adverse scenario, D-SIBs’ CET1 capital adequacy ratio remains above the regulatory minimum of 9 percent (including the capital conservation buffer of 2.5 percent).
- Interbank contagion assessment:
  - 2024 Industry-Wide Stress Test assessed resiliency against cross-border interbank contagion using a tail risk scenario of default of a single country’s banking sector; second-round shocks propagate through counterparty country banking system defaults, funding liquidity stresses from the pullback of funding lines, and asset fire sales.
  - Results suggest that D-SIBs can withstand such shocks in terms of capital and liquidity regulatory requirements.
  - Stress tests draw on the network analysis framework developed by Espinosa-Vega and Solé (2010); an earlier IMF staff analysis under the 2019 FSAP showed Singapore’s banks were most exposed to major advanced economies such as Japan, Hong Kong SAR, and the United States, while shocks originating in or transmitted through Singaporean banks would also affect banks in the ASEAN region.

### FX Funding and Liquidity Resilience
- FX funding indicators:
  - FX loan-to-deposit (LTD) ratio for Singaporean banks declined from above 120 percent in 2019 to around 70 percent in 2024, reaching a level similar to the Singapore dollar LTD ratio.
  - IMF staff analysis shows the FX liquidity coverage ratio (LCR) for three local D-SIBs was close to 100 percent in 2023 and 2024.
  - Overall LCRs for these D-SIBs are well above the regulatory requirement of 100 percent.
- MAS supervisory measures supporting improved resilience:
  - Cashflow-based liquidity stress tests indicating D-SIBs have sufficient FX liquidity buffers on a daily basis, against all currencies including the U.S. dollar.
  - Close engagement with banks to strengthen FX liquidity management and develop contingent FX funding plans.
  - Enhanced monitoring of key funding markets in coordination with central banks globally and efforts to deepen and strengthen Singapore's FX market.
- Data note:
  - Loan to Deposit data covers loans extended by the banking sector, excluding interbank lending; data has a structural break in July 2021 reflecting removal of DBU-ACU divide & revision of MAS 610/1003.

### Conclusions and Policy Implications
- Empirical synthesis:
  - Network analysis and the gravity model point to much stronger links between trade and FDI than between trade and portfolio flows; a reconfiguration of supply chains could lead to redirection of FDI along similar lines.
  - FDI is typically more stable and slow-moving than portfolio flows, but escalating trade measures and prolonged trade policy uncertainty could affect FDI flows.
  - Portfolio flows in Asia have not been associated with trade flows.
- Singapore policy relevance:
  - As Singapore continues to play a significant role in advancing financial integration in Asia, Singaporean banks will remain exposed to cross-border and FX funding risks.
  - Efforts to mitigate these risks will be important for the region’s continued macro-financial stability.

*International Monetary Fund. Singapore: Chapter on banks’ cross-border linkages and vulnerabilities.*

### 1. Gravity Model Results on Trade and Financial Interconnectedness ___________________ 8

### 1. Gravity Model Results on Trade and Financial Interconnectedness

### Introduction
- Paper purpose: take stock of financial interconnectedness in Asia and through Singapore, and explore link between trade and financial flows.
- Key framing:
  - Asian economies are not as financially integrated as their integration in trade would suggest.
  - Asia’s financial integration with the rest of the world has been led by advanced economies in the region, including Singapore.
  - The FDI network closely resembles the trade network; the portfolio assets network is dominated by advanced economies.
  - A gravity-model analysis confirms that countries with strong prior trade linkages are likely to see higher FDI flows but finds no such relationship for portfolio flows.
  - Singaporean banks are exposed to cross-border and FX funding risks; authorities continue mitigation efforts.

### Asia’s Financial Integration with the Rest of the World
- High-level findings:
  - Asia and the Pacific make up over 30 percent of global trade and GDP, hold less than 30 percent of the total inward FDI stock, and make up less than 15 percent of global portfolio liabilities.
  - Asia’s share in global trade grew substantially in the late 2000s and early 2010s but has stabilized since then.
  - Asia’s share of global FDI has risen in recent years; share of global portfolio liabilities has been stable.
- Cross-border positions by country group:
  - Cross-border assets and liabilities (IIP sum) for Asian advanced economies rose from around 100 percent of GDP in 1990 to over 500 percent in 2023.
  - Emerging market economies in the region have cross-border assets and liabilities as percent of GDP that have remained relatively stable since the 1990s.
- Singapore-specific IIP patterns:
  - About half of all FDI coming to Singapore is from either Europe or the United States.
  - The United States has become increasingly important as a source of FDI inflows to Singapore; Europe’s role has diminished.
  - Outbound FDI from Singapore is mostly destined to other parts of Asia.
  - On the portfolio side, about a quarter of portfolio investments is in U.S. assets; China plays an increasing role. Portfolio liabilities are small overall.

### Network Analysis on Financial Integration in Asia
- Method: network analysis using nodes (countries) and edges (flows); eigenvector centrality used to capture influence of nodes.
- Network topologies (2023 snapshots and historical centrality evolution):
  - Trade network:
    - Rise of the Chinese mainland since the 1980s mirrored by decline of Japan.
    - India, Hong Kong SAR, Singapore, ASEAN countries present; Vietnam has risen substantially.
    - Singapore’s centrality has been on a slowly decreasing trend since peaking in the late 1990s.
  - FDI (direct investment) network:
    - Resembles trade network closely.
    - Chinese mainland attracted substantial FDI; Japan’s relative decline observed.
    - Hong Kong SAR and Singapore play larger roles than in the trade network (financial centers).
    - Singapore’s importance has grown steadily, having attracted significant FDI in recent years.
  - Portfolio investment network:
    - Disconnected from trade and FDI networks.
    - Japan remains the biggest Asian holder of portfolio assets, followed by Australia.
    - India and the Chinese mainland are increasing but remain small.
    - Singapore’s centrality broadly stable.
    - Overall, portfolio assets network within Asia is dominated by advanced economies.

### Relationship between Trade and Financial Interconnectedness (Gravity Model)
- Model specification and identification:
  - Modified gravity model where financial exposures (stocks of FDI and portfolio investments) are dependent variables and trade is included on the RHS.
  - Includes origin-time (μi,t), destination-time (δj,t), and origin-destination (φi,j) fixed effects — “fully saturated” specification.
  - Trade variable is lagged by one year to address endogeneity (trade(-1)).
  - Estimation method: maximum likelihood.
  - Data sources: FDI from CDIS, portfolio investments from CPIS, trade from DOTS.
- Main empirical findings:
  - Trade is a key driver for FDI but not for portfolio investments.
  - Baseline regression results (excerpt from Table 1):
    - Trade (-1) coefficients by dependent variable:
      - FDI: 0.005*** (standard error (0.002))
      - Assets: Portfolio: -0.000 (0.001)
      - Assets: Debt: 0.001 (0.002)
      - Assets: Equity: 0.000 (0.001)
      - Liabilities: Portfolio: 0.001 (0.001)
      - Liabilities: Debt: 0.003 (0.002)
      - Liabilities: Equity: 0.002 (0.001)
    - Observations: 143,658; 156,872; 140,455; 119,888; 149,783; 135,842; 114,644 (by column).
    - Country-Year FE: Y (all columns)
    - Counterpart-Year FE: Y (all columns)
    - Country Pair FE: Y (all columns)
    - Pseudo R2: 0.984; 0.993; 0.987; 0.994; 0.993; 0.987; 0.994 (by column).
    - Standard errors clustered by country-pair.
    - Significance: * p<0.10, ** p<0.05, *** p<0.01.
  - Interpretation provided in text:
    - The coefficient for lagged trade is significant at the 99 percent level when using FDI as the dependent variable.
    - A one percentage point increase in the bilateral trade relationship (as percent of average GDP) is associated with an increase in the FDI stock by approximately 0.5 percent.
    - The coefficient for lagged trade is insignificant for portfolio investment (assets/liabilities and debt/equity).
  - Robustness:
    - Findings robust to using shares rather than stocks and to including additional lags of the trade variable.

### Cross-Border Linkages of Singaporean Banks
- Banking sector scale and role:
  - Singapore is an international financial center; banking sector assets totaled 512 percent of GDP in December 2024.
  - Singaporean banks function as intermediaries channeling capital between advanced economies and emerging Asian markets.
- Cross-border asset exposures (December 2024):
  - Banks’ claims on non-residents totaled 116.6 percent of GDP, accounting for about 22 percent of total banking sector assets.
  - Claims composition:
    - Loans and advances including bills financing to non-residents: 67.4 percent of GDP or 12.8 percent of total banking system assets.
    - Securities and equities issued by private non-resident entities: 25.9 percent of GDP or 4.9 percent of total banking system assets.
    - Amounts due from non-resident banks: 23 percent of GDP or 4.4 percent of total banking system assets.
  - Recent change: banks’ claims on non-residents declined from 141.3 percent of GDP in December 2021 to 116.6 percent of GDP in December 2024, mostly due to decline in loans to non-residents from 89.6 percent of GDP in December 2021 to 67.4 percent of GDP in December 2024.
- Funding and liabilities (December 2024):
  - Liabilities of the Singaporean banking system to non-residents totaled 194.4 percent of GDP, accounting for 36.8 percent of total banking sector liabilities.
  - Liability composition:
    - Non-bank deposits by non-residents: 86.7 percent of GDP.
    - Amounts payable to non-resident banks, including interbank loans: 107.7 percent of GDP.
- Structure of local D-SIBs (December 2024 / 2024Q4):
  - Three local D-SIBs (DBS, UOB, and OCBC) accounted for approximately 52 percent of Singapore's banking sector assets in December 2024.
  - D-SIBs have regionally diversified loan portfolios and rely on deposits denominated in foreign currencies.
  - Local D-SIBs’ loans by region (2024Q4):
    - Singapore: 45 percent of total loans.
    - China (including Hong Kong SAR and Taiwan Province of China): 23 percent.
    - Asia and Pacific excluding China: 16 percent.
    - Rest of the world: 16 percent.
  - Currency composition (2024Q4):
    - About 61 percent of local D-SIBs’ loans and deposits were denominated in foreign currencies.
    - Loans denominated in U.S. dollars accounted for about 22 percent of total loans.
    - Deposits denominated in U.S. dollars accounted for about 34 percent of total deposits.
- Market context note:
  - Singapore's foreign exchange derivatives market, the third largest globally, has grown 45 percent since 2019, primarily driven by asset managers, bank treasurers, and corporate treasurers using FX swaps and cross-currency swaps (BIS Triennial Central Bank Survey, 2022).

*International Monetary Fund — FINANCIAL INTERCONNECTEDNESS IN ASIA AND THE PACIFIC—RECENT SHIFTS AND IMPLICATIONS FOR SINGAPORE (June 23, 2025)*

### 18.      Singaporean banks’ extensive cross-border linkages imply distinct vulnerabilities.

### 18.      Singaporean banks’ extensive cross-border linkages imply distinct vulnerabilities.

### Exposures and channels of vulnerability
- Banks in Singapore are exposed to global and regional macroeconomic shocks, a sharp tightening in global financial conditions entailing high global interest rates, shortage in U.S. dollar liquidity, and spillovers from banking sector stresses outside Singapore.  
- Given interconnectedness, these shocks could impact Singaporean banks and broader regional financial stability.  
- Historical resilience cited: Global Financial Crisis, onset of the COVID pandemic, and the banking turmoil in the U.S. and Europe in 2023.  
- Distribution of exposures (as presented): 45% Singapore; 23% China; 16% Rest of Asia-Pacific; 16% Rest of the World.

### MAS stress tests and interbank contagion analysis
- MAS’s Industry-Wide Stress Test in November 2024 assessed D-SIBs under an adverse scenario featuring an intensification of geopolitical conflicts, trade tensions, and global inflationary pressures, entailing a global recession affecting countries and regions on which D-SIBs have exposures, including China and other Asian economies amid elevated interest rates.  
- In the adverse scenario, D-SIBs’ CET1 capital adequacy ratio remains above the regulatory minimum of 9 percent (including the capital conservation buffer of 2.5 percent).  
- The 2024 Industry-Wide Stress Test also assessed resiliency against cross-border interbank contagion using a tail risk scenario of default of a single country’s banking sector; second-round shocks propagate through counterparty country banking system defaults, funding liquidity stresses from the pullback of funding lines, and asset fire sales.  
- Results of the interbank contagion analysis suggest that D-SIBs can withstand such shocks in terms of capital and liquidity regulatory requirements.  
- The stress tests draw on the network analysis framework developed by Espinosa-Vega and Solé (2010); an earlier IMF staff analysis under the 2019 FSAP showed Singapore’s banks were most exposed to major advanced economies such as Japan, Hong Kong SAR, and the United States, while shocks originating in or transmitted through Singaporean banks would also affect banks in the ASEAN region.

### FX funding and liquidity resilience
- FX loan-to-deposit (LTD) ratio for Singaporean banks declined from above 120 percent in 2019 to around 70 percent in 2024, reaching a level similar to the Singapore dollar LTD ratio.  
- IMF staff analysis shows the FX liquidity coverage ratio (LCR) for three local D-SIBs was close to 100 percent in 2023 and 2024.  
- Overall LCRs for these D-SIBs are well above the regulatory requirement of 100 percent.  
- MAS supervisory measures supporting improved resilience include:  
  - Cashflow-based liquidity stress tests indicating D-SIBs have sufficient FX liquidity buffers on a daily basis, against all currencies including the U.S. dollar.  
  - Close engagement with banks to strengthen FX liquidity management and develop contingent FX funding plans.  
  - Enhanced monitoring of key funding markets in coordination with central banks globally and efforts to deepen and strengthen Singapore's FX market.  
- Note: Loan to Deposit data covers loans extended by the banking sector, excluding interbank lending; data has a structural break in July 2021 reflecting removal of DBU-ACU divide & revision of MAS 610/1003.

### Conclusions and policy implications
- Network analysis and the gravity model point to much stronger links between trade and FDI than between trade and portfolio flows; a reconfiguration of supply chains could lead to redirection of FDI along similar lines.  
- FDI is typically more stable and slow-moving than portfolio flows, but escalating trade measures and prolonged trade policy uncertainty could affect FDI flows.  
- Portfolio flows in Asia have not been associated with trade flows.  
- As Singapore continues to play a significant role in advancing financial integration in Asia, Singaporean banks will remain exposed to cross-border and FX funding risks. Efforts to mitigate these risks will be important for the region’s continued macro-financial stability.

*International Monetary Fund. Singapore: Chapter on banks’ cross-border linkages and vulnerabilities.*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2025/english/1sgpea2025002-source-pdf.pdf_