## 1. International Position vs. Foreign Position

## Source details

**Canonical URL:** [1. International Position vs. Foreign Position](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020113-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020113-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020113-print-pdf.pdf.json)

---

### Introduction — role of the US dollar and research questions
- The US dollar (USD) plays a prominent role in global trade and financial flows.
- Pre-crisis buildup: European banks accumulated sizable USD assets financed mainly in short-term wholesale funding markets (repo, commercial paper, certificates of deposits).
- Funding runs in 2007–08 led non-US banks to tap the foreign exchange swap market for USD funding, propagating stress.
- Stable USD deposits outside the United States are typically insufficient to fund global USD credit activities; deployable global USD funding tends to be wholesale, short-term, and volatile.
- Postcrisis regulatory reforms (leverage, capital, liquidity requirements, and the 2016 money market fund reform in the United States) may have tightened the supply of USD funding to non-US banks, increasing reliance on foreign exchange swap markets and nonbank participants.
- Key empirical questions addressed:
  - How USD funding cost responds to identified drivers.
  - Whether USD funding conditions generate financial stress in home economies of non-US global banks and cutbacks in cross-border USD loans to recipient economies.
  - Whether measures of USD funding vulnerability magnify relationships between funding costs, financial stress, and cross-border USD lending.

### Main empirical findings (summary)
- USD funding costs have become more volatile since the global financial crisis, especially for countries relying more on synthetic USD funding.
- Increases in USD funding costs lead to financial stress in home economies (measured by an increase in probability of default of their banking system) and to reduced USD lending from these economies to recipient economies, especially emerging market ones.
- Instrumental variable approaches suggest results are unlikely driven solely by reverse causality.
- The adverse relationship between USD funding costs and both home-economy financial stress and cross-border lending is stronger for home economies with more vulnerable USD funding (greater reliance on foreign currency derivatives, larger liquidity and maturity mismatches).
- Contributions:
  - Confirms determinants of the cross-currency basis (CCB) across a broader set of currencies and shows dependence of CCB relationships on ex-ante reliance on synthetic funding.
  - Provides novel evidence that shocks to USD funding costs can be a source of financial stress for banking systems of home economies.
  - Shows USD funding structure vulnerability amplifies negative effects on cross-border USD lending and on recipient-economy financial stress.
  - Highlights implications of USD dominance for liquidity mismatches on non-U.S. global banks’ balance sheets and financial stability for both lenders and borrowers.

### Stylized facts — scope, data, and positions
- Sample: 26 home economies (17 advanced economies and 9 emerging economies): Australia, Austria, Canada, France, Germany, India, Italy, Japan, Korea, Netherlands, Spain, Sweden, Switzerland, United Kingdom, Brazil, China, Cyprus, Hong Kong SAR, Luxembourg, Malaysia, Mexico, Norway, Russia, Singapore, South Africa, and Turkey.
- Position definitions:
  - International Position (IP): cross-border claims (block A) plus local USD claims of global banks in countries other than the U.S. (block B).
  - Local USD claims of foreign affiliates in the U.S. (block C): positions in U.S. branches (FFIEC002) and in U.S. subsidiaries (FFIEC 031/041 via S&P Market Intelligence).
  - Foreign USD Position (FP): FP = A + B + C; aggregating at home economy level IP + branches + subsidiaries.
- Data caveats:
  - Unable to construct one out of the three fragility measures for 12 economies due to data limitations.
  - For China, Russia, Singapore, Malaysia, Cyprus, and South Africa, BIS data capture only a portion of banks’ worldwide USD positions.

### Evolution of USD activities — key statistics
- FP claims in USD:
  - just over $3 trillion in 2000,
  - rose to $10 trillion before the onset of the crisis,
  - reached more than $12 trillion by 2018.
- Component shares:
  - U.S. branches accounted for close to 19 percent of total FP assets.
  - U.S. subsidiaries accounted for about a tenth of total FP assets.
  - U.S. branches can comprise as much as half or as little as one percent of FP assets.
  - Share of U.S. subsidiaries can vary between zero and 45 percent.
- Shift in nationality distribution:
  - Postcrisis contraction in USD lending by many European banks was offset by increases from Canadian and Japanese banks.
- Share of USD assets in overall banking assets (for the 26 economies):
  - USD assets increased as a share of total banking assets overall.
  - Notable increases in share of USD assets:
    - Canadian banks: increase of five percentage points (pps).
    - Japanese banks: increase of 7 pps.
    - Swedish banks: increase of 2 pps.
  - Aggregate statistic in source fragments: 9.1 percent of USD assets in 2008 to 13.8 percent in 2018 (context preserved as in source).

### Aggregate vulnerability metrics — CCFR, LR, SFR
- Three indicators:
  - Cross-Currency Funding Ratio (CCFR): difference between USD assets and USD liabilities (cross-currency funding gap, CCFG) expressed as a ratio of USD assets; positive CCFR indicates net USD asset position requiring synthetic funding or open FX exposure.
  - USD Liquidity Ratio (LR): ratio of high quality liquid assets (HQLA) to likely USD net cash outflows during a one-month stress scenario.
  - USD Stable Funding Ratio (SFR): ratio of USD deposits, long-term-securities, and long-term swaps to USD loans (ability to secure funding for at least one year).
- Aggregate CCFR findings:
  - Aggregate funding gap amounts to $1.2 trillion.
  - Out of the 26 economies analyzed, 17 registered a positive CCFR in 2018.
  - Almost all of the 26 economies experienced an increase in CCFR since 2012.
- LR and SFR observations:
  - Median LR increased from just over 70 percent in 2010 to over 110 percent in 2018.
  - Larger holdings of HQLA accounted for 30 pps of the increase in the LR.
  - LR could be constructed for 14 economies; virtually all showed increases between 2008 and 2018, with a few European economies and Japan registering small declines since 2016.
  - On average and for most economies, the USD LR lies below a corresponding LR calculated across all currencies using the same methodology.
  - The SFR has remained virtually constant since 2010, with little movement among its components.
  - Appendix sensitivity: assumptions on runoff factors produce sensitivity on the order of 15 pps for the median LR.
  - Note: LR should not be interpreted identically to the regulatory LCR.

### Cross-Currency Basis (CCB) — definition and empirical behavior
- Definition (conceptual): CCB = y$_{$,t,t+n}$ − (y$_{t,t+n}$ − ρ$_{t,t+n}$), where ρ is the forward premium (f − s).
- Interpretation: a positive CCB implies direct USD funding is more costly than synthetic funding; a widening of the basis is defined as the CCB becoming more negative (synthetic cost more expensive relative to direct).
- Empirical behavior:
  - CCB was close to zero prior to the global financial crisis.
  - During the crisis the median widened from close to zero to -100 basis points (bps), and the 25th percentile went from about -5 to -175 bps.
  - As of early February 2018, the median CCB for the sample hovered around -40 to -50 bps, indicating persistent deviation from covered interest parity (CIP).

### Drivers of CCB and econometric framework
- Monthly panel estimations January 2000 – March 2018 link CCB to drivers X_{i,t} including:
  - VIX (U.S. market volatility)
  - LIBOR-OIS spread
  - Term spread differential between the U.S. and currency i (10-year minus 3-month government yields)
  - Bid-ask spread for the spot exchange rate
  - Implied FX option volatility for currency i
  - Average default probability of banks with home currency i (EDF from Moody’s Analytics, one-year horizon)
  - Broad USD index (DXY)
  - CCFR (cross-currency funding gap)
- Baseline results:
  - Drivers have expected signs and are statistically significant.
  - CCFR is positively associated with a widening of the basis (higher CCFR → more negative CCB).
  - Results robust across alternative CCB measures based on different tenors and/or benchmark rates.
- Interaction tests:
  - Econometric specifications include interaction terms between CCFR and each driver to test amplification.
  - CCFR amplifies effects of CCB drivers: impact of a given driver on the CCB is greater when the banking system has a larger gap between USD assets and liabilities.

### Quantitative coefficients (selected standardized results from drivers table)
- CCFR standardized coefficients (selected):
  - 3m CCB (LIBOR): -0.096***
  - 1m CCB (LIBOR): -0.252***
  - 1m CCB (OIS): -0.307***
- VIX standardized coefficients:
  - 3m CCB (LIBOR): 0.100***
  - 1m CCB (LIBOR): 0.102***
- FX implied volatility standardized coefficients:
  - 3m CCB (LIBOR): -0.169***
  - 1y CCB (LIBOR): -0.062***
- Interaction examples (CCFR x driver, standardized):
  - LIBOR-OIS spread x CCFR (3m CCB LIBOR): -0.124***
  - VIX x CCFR (3m CCB LIBOR): 0.308***
  - Term spread differential x CCFR (3m CCB LIBOR): 0.174***
- Effects at low vs high CCFR (standardized):
  - Term spread differential: Low CCFR 0.275; High CCFR 0.510
  - VIX: Low CCFR 0.069; High CCFR 0.515
  - LIBOR-OIS spread: Low CCFR -0.016; High CCFR -0.217

### Effects of regulation and market structure on the basis
- Tests for discrete jumps around major regulatory reforms using dummies for:
  - stress-VaR (January–December 2013)
  - supplementary leverage ratio (January–December 2014)
  - liquidity coverage ratio (January–December 2015)
  - money market mutual fund (MMMF) reform (January–December 2016)
- Findings:
  - Liquidity coverage ratio introduction and MMMF reform appear associated with heightened sensitivity of the basis to demand for synthetic USD funding (as proxied by CCFR).
  - The highest sensitivity occurs around MMMF reform in 2016, when funding drained out of prime institutional money market funds, reducing wholesale dollar lending and inducing foreign banks to resort to synthetic dollar funding, widening the basis.
- Literature explanations for persistent CIP violations include imbalances between USD investment opportunities and funding supply (related to CCFR), post-crisis regulatory changes, FX swap market structure changes, interest-rate differentials, USD liquidity needs of non-US banks, and credit risk considerations.

### Robustness and endogeneity checks
- Lagging CCFR yields similar results (addresses reverse causality concerns).
- Unrestricted panel VAR estimated and orthogonalized impulse response functions reported.
- Granger causality tests on CCFR and CCB find CCFR Granger causes CCB.
- Instrumental-variables approach: two-stage least-squares using U.S. monetary policy shocks (Federal Funds rate shocks and policy news shocks) as instruments for CCB; second-stage results show the positive relationship between instrumented basis and home-economy financial stress remains statistically significant.

### Transmission to home-economy financial stress — core results
- Baseline regression links changes in home-economy banking-sector stress (ΔY_{i,t} measured by one-year ahead probability of default, PD) to changes in (minus) the CCB (so an increase corresponds to widening of the CCB).
- Controls: interest-rate differentials with U.S., home inflation, log USD nominal effective exchange rate, level and volatility of log VIX, home real GDP growth, banking-sector controls (ratios of capital, deposits, net loans to assets, cost-to-income ratio); controls are moving averages from quarter t-4 to t-1.
- Key findings:
  - β is positive and statistically significant for PD: increases in USD funding costs are associated with greater financial stress in the home economy.
  - Evidence of nonlinearity: association strengthens for larger increases in costs.
  - Global financial crisis (2007–09) and European debt crisis (2011–12) exhibit particularly strong relationships between USD funding costs and home-economy financial stress.
- Instrumental-variable two-stage least-squares with U.S. monetary policy shocks confirms the positive relationship remains statistically significant.

### Quantified amplification (illustrative magnitudes)
- For an economy with a “low” CCFR (first quintile), a 50 basis-point increase in the CCB has a negligible effect on the probability of default.
- For an economy with a “high” CCFR (fourth quintile), a 50 basis-point increase in the CCB is equal to 0.41 standard deviations in PD, or a 14 percent increase.
- Given the average quarterly increase in PD at the height of the crisis was 34 percent, this amplification could account for about two-fifths of the increase in financial stress.
- When either LR or SFR is low, amplification rises to 0.4 standard deviations.

### Mitigating factors for home-economy transmission
- Swap lines:
  - Swap lines between the US Federal Reserve and other central banks appear to mitigate the international transmission of USD funding shocks.
  - When a swap line was announced or in effect, the relationship between USD funding costs and home economy financial stress is weak and statistically insignificant.
  - Peak number of central banks with temporary USD swap arrangements: 14 in October 2008; stabilized to five major advanced economy central banks in May 2010 with full allotment.
- Banking-sector fundamentals:
  - Stronger bank capital (capital-asset ratio), profitability (return on assets), and liquidity (cash-asset ratio) weaken the relationship between USD funding costs and home economy financial stress.
  - Example magnitudes:
    - A 50 basis-point increase in USD funding costs is associated with a 0.40 standard deviation (14 pps) increase in the probability of default when the capital ratio is low by historical standards.
    - The increase is 0.25 standard deviations (8 pps) when the capital ratio is high.
- International reserves:
  - Higher reserves mitigate amplification from USD funding vulnerability; triple-interaction terms indicate that greater reserve holdings weaken the amplification from USD fragility measures.

### Effects on cross-border USD lending — main findings
- Baseline bilateral regression (home i → recipient j, log USD claims averaged t+1 to t+4, multiplied by 100) finds 훽 < 0: higher USD funding costs → cutback in USD lending.
- Main quantitative results:
  - A 50 bps increase in USD funding costs is associated with a 5 percent decline in USD lending.
  - Cutbacks are stronger when the home economy is an EM or the recipient economy is an EM.
  - Nonlinearities: funding-cost quadratic terms show stronger effects for larger shocks.
- Substitution and incidence:
  - Some substitution from other USD cross-border lenders is observed; estimated substitution is partial, about half of the lost financing (example computation: θ ≈ 51.8 percent).
  - For EM recipients, substitution across USD cross-border lenders is more limited: estimated θ ≈ 46.6 percent.
  - Total USD lending (cross-border + domestic) after broad-based funding shock falls: Column 3 finds total USD lending reduced by 2.8 percent after a broad-based funding shock.
  - Substitution across currencies:
    - For advanced-economy recipients, substitution across currencies is noticeably greater and may fully compensate for loss in USD cross-border lending.
    - For EM borrowers, compensation across currencies does not occur.

### Amplification of lending cuts by home-economy vulnerability
- Expanded specification interacts USD funding cost shocks with measures of USD funding vulnerability (CCFR, LR, SFR, USD HQLA-to-assets).
- Results:
  - All three measures amplify the cutback in USD lending associated with increased USD funding costs.
  - Example magnitude:
    - A 50 basis-point increase in USD funding costs is statistically insignificant when CCFR is low (first quintile), but increases to 5.82 percent (statistically significant) when CCFR is high (fourth quintile).

### Mitigating factors for lending transmission
- Bank fundamentals:
  - High-liquidity banking sectors do not cut back USD lending following a 50 bps increase in USD funding costs; low-liquidity systems cut back by about 3 percent.
- Swap lines:
  - Home economies with a swap line with the US Federal Reserve are not likely to cut back USD lending; those lacking such arrangements are estimated to cut back by 5.6 percent on average.
- International reserves:
  - The estimated cutback is mitigated by 2.9 percentage points if domestic central bank holdings of international reserves are higher by 10 percentage points of GDP.
- Effects persist when controlling for CCFR and its interactions.

### Recipient economy financial stress and concentration
- Specification linking recipient country PD or FCI to weighted average changes in USD funding costs across lenders (weights are lenders’ share in recipient j’s USD lending).
- Results:
  - Adverse effect of USD funding costs on recipient economy financial stress (both banking probability of default and financial conditions).
  - Effect concentrated in the top ten recipients of USD cross-border lending.

### Policy implications and recommendations
- Monitor USD funding fragility:
  - Regulators should monitor USD funding fragility of local banks and develop/enhance currency-specific liquidity risk frameworks, including stress tests, emergency funding strategies, and resolution planning.
  - Monitoring tools: cross-currency funding ratio, liquidity ratio, and stable funding ratio measures introduced in the chapter.
- Reserves and swap lines:
  - Assess reserve adequacy taking into account the stabilizing role of international reserves against USD funding market stress.
  - Access to USD liquidity through swap lines can contribute to stability, including through signaling effects.
- Global financial safety net:
  - Consider strengthening the global financial safety net, including adequate IMF resources such as flexible credit lines.
- Structural considerations:
  - Postcrisis regulatory reforms improved resilience but may have had unintended effects tightening USD funding supply to non-US banks; policy design should weigh cross-border USD funding implications.

### Research note
- Analysis is at the banking sector level due to broad country coverage; testing the mechanism at granular, bank-level data is an important area for further research.

*Source: wpiea2020113-print-pdf*

### 1. International Position vs. Foreign Position .........................................................................

### 1. International Position vs. Foreign Position

### Introduction
- The US dollar (USD) plays a prominent role in global trade and financial flows.
- Before the global financial crisis, European banks accumulated sizable USD assets financed mainly in short-term wholesale funding markets (repo, commercial paper, certificates of deposits).
- When these markets became impaired in 2007–08, non-US banks tapped the foreign exchange swap market to obtain USD funding, propagating stress through this market.
- Non-US banks remain significant intermediaries of USD transactions; USD intermediation provides global liquidity allocation benefits but can create foreign currency liquidity mismatches for non-US banks.
- Stable USD deposits outside the United States are typically insufficient to fund global USD credit activities; USD funding sources that are deployable globally tend to be wholesale, short-term, and volatile.
- Postcrisis regulatory reforms (leverage, capital, liquidity requirements, and the 2016 money market fund reform in the United States) may have tightened the supply of USD funding to non-US banks, increasing reliance on foreign exchange swap markets and nonbank participants.
- Greater USD funding fragility—reflected in liquidity and maturity mismatches or reliance on short-term funding—can amplify shocks to USD funding costs, affecting banks’ financial stress and global credit supply.
- In early 2020, the cross-currency basis (proxy for marginal cost of USD funding outside the U.S.) spiked across major currencies, with enlargements for some currencies exceeding those in the global financial crisis; the Federal Reserve enhanced and extended swap lines to fourteen central banks in late March, after which bases largely reverted to pre-2020 levels by June.
- This paper investigates:
  - How USD funding cost responds to identified drivers.
  - Whether USD funding conditions generate financial stress in home economies of non-US global banks and cutbacks in cross-border USD loans to recipient economies.
  - Whether measures of USD funding vulnerability magnify relationships between funding costs, financial stress, and cross-border USD lending.
- Main empirical findings summarized here:
  - USD funding costs have become more volatile since the global financial crisis, especially for countries relying more on synthetic USD funding.
  - Increases in USD funding costs lead to financial stress in home economies (measured by an increase in probability of default of their banking system) and to reduced USD lending from these economies to recipient economies, especially emerging market ones.
  - Instrumental variable approaches suggest results are unlikely driven solely by reverse causality.
  - The adverse relationship between USD funding costs and both home-economy financial stress and cross-border lending is stronger for home economies with more vulnerable USD funding (greater reliance on foreign currency derivatives, larger liquidity and maturity mismatches).
- Contributions:
  - Confirms determinants of the cross-currency basis (CCB) across a broader set of currencies and shows dependence of CCB relationships on ex-ante reliance on synthetic funding.
  - Provides novel evidence that shocks to USD funding costs can be a source of financial stress for banking systems of home economies.
  - Shows USD funding structure vulnerability amplifies negative effects on cross-border USD lending and on recipient-economy financial stress.
  - Highlights implications of USD dominance for liquidity mismatches on non-U.S. global banks’ balance sheets and financial stability for both lenders and borrowers.

*Structure note from source*: Section II reviews stylized facts; Section III analyzes drivers and effects of regulatory and policy changes; Section IV reviews association between USD funding costs and home economy financial stress; Section V does so for recipient economies; Section VI concludes.

### Stylized Facts — Data and Concepts
- Data sources:
  - BIS locational banking statistics (LBS) on a nationality basis for 26 economies that are home to major global non-US banks, used to construct USD funding fragility measures.
  - Aggregation of international position (IP) with positions in U.S. branches and U.S. subsidiaries yields the foreign USD position (FP = A + B + C).
- Definitions of positions (as used in the paper):
  - International Position (IP): cross-border claims (block A) plus local USD claims of global banks in countries other than the U.S. (block B).
  - Local USD claims of foreign affiliates in the U.S. (block C): positions in U.S. branches (FFIEC002) and in U.S. subsidiaries (FFIEC 031/041 via S&P Market Intelligence).
  - Foreign USD Position (FP): FP = A + B + C; aggregating at home economy level IP + branches + subsidiaries.
- Sample selection criteria for the 26 home economies:
  - Existence of cross-currency swaps for the relevant currency;
  - Availability of BIS LBS data starting in 2015 at the latest; and either:
    - global importance: international USD claims of at least 50 billion; or
    - domestic importance: share of international claims in USD of at least 5 percent of the country's total international claims plus domestic banks' local claims in local currency.
- The resulting sample includes seventeen advanced economies and nine emerging economies:
  - Australia, Austria, Canada, France, Germany, India, Italy, Japan, Korea, Netherlands, Spain, Sweden, Switzerland, United Kingdom, Brazil, China, Cyprus, Hong Kong SAR, Luxembourg, Malaysia, Mexico, Norway, Russia, Singapore, South Africa, and Turkey.
- Data caveats:
  - Unable to construct one out of the three fragility measures for 12 of the economies due to data limitations.
  - For China, Russia, Singapore, Malaysia, Cyprus, and South Africa, BIS data capture only a portion of banks’ worldwide USD positions and therefore may not be comprehensive.

### Stylized Facts — Measuring US Dollar Operations and Funding of Global Non-US Banks
- Evolution of USD activities:
  - FP claims in USD:
    - just over $3 trillion in 2000,
    - rose to $10 trillion before the onset of the crisis,
    - reached more than $12 trillion by 2018.
  - All three components (IP, U.S. branches, U.S. subsidiaries) contributed to the increase.
  - U.S. branches accounted for close to 19 percent of total FP assets.
  - U.S. subsidiaries accounted for about a tenth of total FP assets.
  - Relative reliance varies substantially by home economy:
    - U.S. branches can comprise as much as half or as little as one percent of FP assets.
    - Share of U.S. subsidiaries can vary between zero and 45 percent.
- Shift in home nationality distribution:
  - Postcrisis contraction in USD lending by many European banks was offset by increases from Canadian and Japanese banks, with substantial and relatively uninterrupted increases in USD assets for those countries.
- Share of USD assets in overall banking assets:
  - For the group of 26 economies, USD assets have increased as a share of total banking assets overall.
  - Increases occurred in all but five home economies analyzed.
  - Notable increases in share of USD assets:
    - Canadian banks: increase of five percentage points (pps).
    - Japanese banks: increase of 7 pps.
    - Swedish banks: increase of 2 pps.

### Dollar Funding Vulnerability Indicators
- Three indicators used to quantify vulnerability to USD funding shocks:
  - Cross-Currency Funding Ratio (CCFR)
  - USD Liquidity Ratio (LR)
  - USD Stable Funding Ratio (SFR)
- CCFR definition and interpretation:
  - CCFR is defined as the difference between USD assets and USD liabilities (the cross-currency funding gap, CCFG), expressed as a ratio of USD assets.
  - A positive CCFR indicates a net asset position in USD, implying some USD assets are not directly funded through USD liabilities.
  - The gap is either synthetically funded via other currencies transformed to USD using derivatives (FX swaps or cross-currency swaps) or held as a net open foreign currency position (the latter is typically limited due to regulatory constraints).
  - Larger CCFR values signal higher vulnerability to shocks to USD funding costs, particularly when synthetic funding relies on short-term FX swaps or swaps with tenors shorter than USD assets.
  - Synthetic funding is often used when cheaper than direct USD issuance, and banks with limited access to insured retail deposits may rely more heavily on wholesale USD funding.
  - In stress periods, difficulties obtaining USD funding can push banks to seek synthetic funding or reduce USD lending.
- Origins of indices and methodology notes:
  - Cross-currency funding gap initially constructed in McGuire and von Peter (2009).
  - LR and SFR initially constructed in Chapter 1 of the IMF’s April 2018 Global Financial Stability Report.
  - The three vulnerability measures rely on BIS unpublished restricted LBS data on a nationality basis; see Appendix 1 in the source for construction details.
  - Figures showing USD vulnerability indicators report the IP + B definition of USD activities (International Position plus branches in the U.S., excluding U.S. subsidiaries), under the presumption that liquidity or funding of U.S. subsidiaries cannot be easily transferred to the parent due to regulatory constraints.
  - Robustness checks show no significant change in results when moving from IP + B to FP definition (which includes subsidiaries).
- Observed trends:
  - Together with expansion in USD activities, the CCFR has increased for many economies since the crisis.
  - Figure references in source: Figure 1 shows position definitions; Figure 2 displays FP evolution and component shares; Figure 3, panel 1 displays aggregate CCFG and CCFR for a sample of 13 countries since 2000 (note: some economies enter the sample after 2000, e.g., China and Russia enter in 2015, accounting for a discrete jump in USD claims).

### Key implications and mechanisms emphasized in the source
- Reliance on synthetic USD funding increases exposure to liquidity and rollover risks when synthetic funding is short-term or mismatched with asset tenors.
- Regulatory and market structure changes can reduce the supply of wholesale USD funding to non-US banks, increasing reliance on FX swap markets and potentially nonbank suppliers whose behavior in stress is uncertain.
- USD funding cost volatility can transmit to:
  - Increases in the probability of default of home-economy banking systems.
  - Reductions in cross-border USD lending from home economies, particularly to emerging-market recipient economies.
- Amplification effects:
  - Home economies with more vulnerable USD funding structures (higher CCFR, lower LR, lower SFR) experience stronger negative effects from USD funding cost shocks on financial stress and cross-border lending.
- Policy-relevant observations:
  - Central bank swap lines (as used by the Federal Reserve in March 2020) can be effective in arresting USD funding strains.
  - Postcrisis regulatory reforms improved resilience but may have had unintended effects tightening USD funding supply to non-US banks.

*Source: IMF Working Paper (content unit: "1. International Position vs. Foreign Position" from the provided PDF).*

### 9.1 percent of USD assets in 2008 to 13.8 percent in 2018. For these economies, t he

### wpiea2020113-print-pdf - 9.1 percent of USD assets in 2008 to 13.8 percent in 2018. For these economies, t he

### Aggregate USD exposure and CCFR findings
- Aggregate funding gap amounts to $1.2 trillion.
- Out of the 26 economies analyzed, 17 registered a positive CCFR in 2018.
- Almost all of the 26 economies experienced an increase in CCFR since 2012.
- The CCFR captures the gap between USD assets and USD liabilities and is used as an indicator of banking-system vulnerability to USD funding stress.

### USD liquidity and stable funding metrics (LR and SFR)
- LR (USD liquidity ratio) is defined as the ratio of high quality liquid assets (HQLA) to likely USD net cash outflows during a one-month stress scenario.
- SFR (stable funding ratio) is defined as the ratio of USD deposits, long-term-securities, and long-term swaps to USD loans (ability to secure funding for at least one year).
- Median LR increased from just over 70 percent in 2010 to over 110 percent in 2018.
- Larger holdings of HQLA accounted for 30 pps of the increase in the LR.
- LR could be constructed for 14 economies; virtually all showed increases between 2008 and 2018, with a few European economies and Japan registering small declines since 2016.
- On average and for most economies, the USD LR lies below a corresponding LR calculated across all currencies using the same methodology.
- The SFR has remained virtually constant since 2010, with little movement among its components.
- Appendix sensitivity: assumptions on runoff factors produce sensitivity on the order of 15 pps for the median LR.
- Data limitation note: assumptions on swap maturity share required; LR should not be interpreted identically to the regulatory LCR.

### Cross-Currency Basis (CCB): definition and empirical behavior
- CCB is defined as the difference between direct USD funding cost and the synthetic USD funding cost (via domestic currency borrowing plus FX forward).
- Formula (conceptual): CCB = y$_{$,t,t+n}$ − (y$_{t,t+n}$ − ρ$_{t,t+n}$), where ρ is the forward premium (f − s).
- Interpretation: a positive CCB implies direct USD funding is more costly than synthetic funding; a widening of the basis is defined as the CCB becoming more negative (synthetic cost more expensive relative to direct).
- The CCB was close to zero prior to the global financial crisis; during the crisis the median widened from close to zero to -100 basis points (bps), and the 25th percentile went from about -5 to -175 bps.
- As of early February 2018, the median CCB for the sample hovered around -40 to -50 bps, indicating persistent deviation from covered interest parity (CIP).

### Drivers of CCB and econometric specification
- Monthly panel estimations for January 2000 – March 2018 link CCB to drivers X_{i,t} including:
  - VIX (U.S. market volatility)
  - LIBOR-OIS spread
  - Term spread differential between the U.S. and currency i (10-year minus 3-month government yields)
  - Bid-ask spread for the spot exchange rate
  - Implied FX option volatility for currency i
  - Average default probability of banks with home currency i (EDF from Moody’s Analytics, one-year horizon)
  - Broad USD index (DXY)
  - CCFR (cross-currency funding gap)
- Econometric specifications include interaction terms between CCFR and each driver to test amplification.
- Baseline results: drivers have expected signs and are statistically significant; CCFR is positively associated with a widening of the basis (higher CCFR → more negative CCB).
- Results are robust across alternative CCB measures based on different tenors and/or benchmark rates.

### Amplification role of CCFR and other vulnerabilities
- CCFR amplifies effects of CCB drivers: impact of a given driver on the CCB is greater when the banking system has a larger gap between USD assets and liabilities.
- Amplification illustrated by comparing low (first quintile) versus high (fourth quintile) CCFR histories for each economy; for some drivers (e.g., VIX, LIBOR-OIS spread), amplification can be large.
- Other vulnerability indicators (USD asset share, LR, SFR) also act as amplifiers: lower USD liquidity (low LR) or less stable funding (low SFR) strengthens transmission from USD funding costs to banking-sector stress.

### Effects of regulation and market structure
- Regression variant tests discrete jumps around dates of major regulatory reforms using four dummy variables: stress-VaR (January–December 2013), supplementary leverage ratio (January–December 2014), liquidity coverage ratio (January–December 2015), and money market mutual fund (MMMF) reform (January–December 2016).
- Liquidity coverage ratio introduction and MMMF reform appear associated with heightened sensitivity of the basis to demand for synthetic USD funding (as proxied by CCFR).
- The highest sensitivity occurs around MMMF reform in 2016, when funding drained out of prime institutional money market funds, reducing wholesale dollar lending and inducing foreign banks to resort to synthetic dollar funding, widening the basis.
- Literature explanations for persistent CIP violations include: persistent imbalance between USD investment opportunities and funding supply (related to CCFR), post-crisis regulatory changes increasing direct USD funding costs and reducing FX swap market USD supply, changes in FX swap market structure (reduced bank role, increased real money investor participation), interest-rate differentials, USD liquidity needs of non-US banks, wealth of U.S. arbitrageurs, and credit risk considerations.

### Robustness and endogeneity checks
- Addressed potential reverse causality (CCB affecting banks’ USD asset positions) by lagging CCFR—results similar.
- Estimated an unrestricted panel VAR treating variables as jointly endogenous; orthogonalized impulse response functions reported (Figure A.1 in source).
- Granger causality tests on CCFR and CCB (panel VAR with two equations) find CCFR Granger causes CCB.
- Instrumental-variables approach: two-stage least-squares using U.S. monetary policy shocks (Federal Funds rate shocks and policy news shocks per Nakamura and Steinsson (2018)) as instruments for CCB; second-stage results show the positive relationship between instrumented basis and home-economy financial stress remains statistically significant.

### Transmission to home-economy financial stress
- Baseline regression links changes in home-economy banking-sector stress (ΔY_{i,t} measured by one-year ahead probability of default, PD) to changes in (minus) the CCB (so an increase corresponds to widening of the CCB).
- Controls include interest-rate differentials with U.S., home inflation, log USD nominal effective exchange rate, level and volatility of log VIX, home real GDP growth, and banking-sector controls (ratios of capital, deposits, net loans to assets, cost-to-income ratio). Controls expressed as moving averages from quarter t-4 to t-1.
- Key findings:
  - β is positive and statistically significant for PD: increases in USD funding costs are associated with greater financial stress in the home economy.
  - Evidence of nonlinearity: association strengthens for larger increases in costs.
  - Global financial crisis (2007–09) and European debt crisis (2011–12) exhibit particularly strong relationships between USD funding costs and home-economy financial stress.
- Two-stage least-squares with U.S. monetary policy shocks as instruments confirms the positive relationship remains statistically significant.

### Quantified amplification effects (illustrative magnitudes)
- For an economy with a “low” CCFR (first quintile), a 50 basis-point increase in the CCB has a negligible effect on the probability of default.
- For an economy with a “high” CCFR (fourth quintile), a 50 basis-point increase in the CCB is equal to 0.41 standard deviations in PD, or a 14 percent increase.
- Given the average quarterly increase in PD at the height of the crisis was 34 percent, this amplification could account for about two-fifths of the increase in financial stress.
- Amplification effects are not economically significant when LR or SFR are high (fourth quintile) but rise to 0.3 – [text truncates in source at this point].

*Source: wpiea2020113-print-pdf*

### 0.4 standard deviations when either ratio is low.

### wpiea2020113-print-pdf - 0.4 standard deviations when either ratio is low.

### Identification, robustness, and key quantitative mechanics
- Identification strategy: difference-in-difference comparing changes in the cost of USD funding and financial stress across periods and countries with different degrees of USD funding vulnerability.
- Euro area subsample robustness:
  - Results hold in the Euro area subsample (eight economies): an increase in the basis is associated with increased financial stress in the home economy.
  - Results remain when excluding Germany and France (Table A.2).
- Numerical illustration (from Table 7, Column (3) coefficients):
  - Cross-currency basis coefficient: -0.066.
  - Interaction with CCFR coefficient: 0.401.
  - 20th and 80th percentiles of positive CCFRs: 0.18 and 0.83 respectively.
  - Impact of one-standard deviation increase in the cross-currency basis (32.49) on home country financial stress:
    - 20th percentile: 0.0065 = -0.066 + 0.401×0.18.
    - 80th percentile: 0.267 = -0.066 + 0.401×0.83.
  - Impact of a 50 bp increase in the cross-currency basis (multiply by 50/32.49):
    - 20th percentile: 0.01.
    - 80th percentile: 0.41.
  - Translation into probability of default: 0.41×0.337 = 13.8 percent.
- Robustness checks:
  - Results include home economy fixed effects and time-varying macro and banking controls.
  - Results persist when adding time fixed effects to control for global trends.

### Amplifying and mitigating factors (home economy)
- Swap lines:
  - Swap lines between the US Federal Reserve and other central banks appear to mitigate the international transmission of shocks to USD funding costs.
  - When a swap line was announced or in effect, the relationship between USD funding costs and home economy financial stress is weak and statistically insignificant (Table 8, Columns 1–4).
  - Peak number of central banks with temporary USD swap arrangements: 14 in October 2008; stabilized to five major advanced economy central banks in May 2010 with full allotment.
- Banking-sector fundamentals:
  - Stronger bank capital (capital-asset ratio), profitability (return on assets), and liquidity (cash-asset ratio) weaken the relationship between USD funding costs and home economy financial stress (Table 9).
  - Economic magnitudes:
    - A 50 basis-point increase in USD funding costs is associated with a 0.40 standard deviation (14 peps) increase in the probability of default when the capital ratio is low by historical standards.
    - The increase is 0.25 standard deviations (8 peps) when the capital ratio is high.
- International reserves:
  - Regressions (specification (5)) test for mitigating role of international reserves; expected signs: 훽21 > 0 and 훽22 < 0.
  - Table 10: lower liquidity and stable funding amplify the USD funding–financial stress relationship; greater reserve holdings weaken this amplification.
  - Regression evidence: higher reserves mitigate amplification from USD funding vulnerability.
- Summary implication:
  - Conditional on USD funding fragility, tightening USD funding conditions are transmitted to financial stress in home economies of non-US global banks; swap lines, stronger bank fundamentals, and larger reserves mitigate transmission.

### USD funding costs and cross-border USD lending (recipient effects)
- Baseline empirical specification relating changes in USD funding costs to cross-border USD claims (equation (6)):
  - Dependent variable: log of USD claims from home economy i to recipient j, averaged over quarters t+1 to t+4 (multiplied by 100).
  - Main test: 훽 < 0 indicates higher USD funding costs → cutback in USD lending.
- Main results (Table 11):
  - Increases in USD funding costs are followed by a significant cutback in USD lending.
  - A 50 bps increase in USD funding costs is associated with a 5 percent decline in USD lending.
  - Cutbacks are stronger when the home economy is an EM (Column 2) or the recipient economy is an EM (Column 3).
  - Nonlinearities: quadratic term in USD funding cost (Column 4) and interaction with initial level of USD funding costs (Column 5).
  - Seemingly Unrelated Regression results (Columns 6 and 7) are similar to fixed-effects regressions.

### Degree of substitutability of USD lending (how recipients offset cutbacks)
- Framework (equations (7), (7a), (7b)):
  - Examine whether recipients substitute lost lending by (i) borrowing more USD from other cross-border lenders, (ii) borrowing more USD domestically, or (iii) borrowing more in other currencies cross-border.
  - Substitution coefficient 휃 = −(훽1/β) ×25; interpretation:
    - θ = 1: full substitution (all other 25 lenders fully offset cutback).
    - θ < 1: partial substitution.
    - θ = 0: no substitution.
- Empirical findings (Table 12):
  - Some substitution from other USD cross-border lenders: 훽1 positive and significant for USD cross-border lending (Column 1).
  - Estimated substitution is partial, about half of the lost financing:
    - Computation example: θ = −(0.113×25)/5.31 ≈ 51.8%.
  - For EM recipients, substitution possibilities are more limited:
    - Estimated degree of substitution across USD cross-border lenders θ of 46.6 percent (from Column 2); computed as θ = −((0.132 −0.033)×25)/5.31 ≈46.6%.
  - Total USD lending (cross-border + domestic) after broad-based funding shock falls:
    - Column 3: total USD lending reduced by 2.8 percent after a broad-based funding shock.
  - Substitution across currencies:
    - For AEs, substitution across currencies is noticeably greater.
    - Column 5: 훽1 negative but not significant for substitution across currencies, implying cross-border lending in other currencies appears to fully compensate for loss in USD cross-border lending for AEs.
    - For EM borrowers (Column 6), this compensation across currencies does not occur.

### Home-economy funding vulnerability amplifying lending cuts
- Expanded specification (equation (8)) interacts USD funding cost shocks with measures of USD funding vulnerability (AMP): within-country quintiles of CCFR, LR, SFR, and ratio of USD HQLA to USD assets.
- Results (Table 13, Columns 1–3):
  - All three measures of vulnerability amplify the cutback in USD lending associated with increased USD funding costs.
  - Example magnitude:
    - A 50 basis-point increase in USD funding costs is statistically insignificant when CCFR is low (first quintile), but increases to 5.82 percent (statistically significant) when CCFR is high (fourth quintile).

### Amplifying or mitigating effects of other factors on lending transmission
- Regressions (Table 14) examine whether bank fundamentals, swap lines, and reserves mitigate effects of USD funding cost increases on cross-border lending.
- Main findings:
  - Bank fundamentals:
    - High-liquidity banking sectors do not cut back USD lending following a 50 bps increase in USD funding costs; low-liquidity systems cut back by about 3 percent.
  - Swap lines:
    - Global non-US banks from economies with a swap line with the US Federal Reserve are not likely to cut back USD lending; those lacking such arrangements are estimated to cut back by 5.6 percent on average.
  - International reserves:
    - The estimated cutback is mitigated by 2.9 percentage points if domestic central bank holdings of international reserves are higher by 10 percentage points of GDP.
  - Columns 5–8 control for USD funding vulnerability (CCFR and interaction) and confirm the mitigating roles of fundamentals, swap lines, and reserves.

### Recipient economy financial stress and concentration
- Specification linking recipient country PD or FCI to weighted average changes in USD funding costs across lenders (equation (9)):
  - Weighting: R_i,j,t is the share of lending economy i in USD lending received by economy j in quarter t.
- Results (Table 15):
  - There is an adverse effect of USD funding costs on recipient economy financial stress (both banking probability of default and financial conditions).
  - Effect concentrated in the top ten recipients of USD cross-border lending.

### Conclusion and policy implications (explicit findings and recommendations)
- Stylized facts and trends:
  - USD operations of global non-US banks have been increasing steadily post-crisis, including cross-border operations, operations in branches outside the U.S., and in branches and subsidiaries in the U.S.
  - For some economies, USD operations are increasing as a share of domestic banking system assets.
  - The gap between USD assets and liabilities (CCFR) has widened since the GFC for many economies.
  - USD liquidity and stable funding ratios have improved post-crisis but remain low in many economies in USD terms.
  - Post-crisis the cross-currency basis (CCB) has not reverted to low pre-crisis levels.
- Core empirical conclusions:
  - Increases in USD funding costs produce stress in the domestic financial system of economies that are home to global non-US banks.
  - The effect is amplified by larger CCFR and weakened when USD liquidity or stable funding is stronger, and when domestic banking system health is stronger.
  - Increases in USD funding costs are associated with negative spillovers to recipients of cross-border lending by non-US banks; EM borrowers are particularly susceptible.
  - Swap line arrangements and central bank international reserves can mitigate some adverse effects.
- Policy recommendations and implications:
  - Regulators should monitor USD funding fragility of local banks and develop/enhance currency-specific liquidity risk frameworks, including stress tests, emergency funding strategies, and resolution planning.
  - Monitoring tools: cross-currency funding ratio, liquidity ratio, and stable funding ratio measures introduced in the chapter.
  - Assess reserve adequacy taking into account the stabilizing role of international reserves against USD funding market stress.
  - Access to USD liquidity through swap lines can contribute to stability, including through signaling effects.
  - Consider strengthening the global financial safety net, including adequate IMF resources such as flexible credit lines.
- Research note:
  - The analysis is at the banking sector level due to broad country coverage; testing the mechanism at granular, bank-level data is an important area for further research.

*Source: wpiea2020113-print-pdf - 0.4 standard deviations when either ratio is low.*

### References

### References

### Bibliographic entries
- Aldasoro, Iñaki. Thorsten Ehlers, Egemen Eren, and Robert N. McCauley. 2017. “Non-US Banks’ Global Dollar Funding Grows Despite US Money Market Reform.” BIS Quarterly Review (March): 22-23.
- Aldasoro, Iñaki. Thorsten Ehlers, and Egemen Eren. 2018. “Global Banks, Dollar Funding, and Regulation. 2018. BIS Working Paper No.708.
- Aldasoro, Iñaki, and Torsten Ehlers. 2018. “The Geography of Dollar Funding of Non-US Banks.” BIS Quarterly Review (December): 15–26.
- Avdjiev, Stefan, Wenxin Du, Catherine Koch, and Hyun Song Shin. 2019. “The Dollar, Bank Leverage and the Deviation from Covered Interest Parity.” American Economic Review: Insights. 1(2): 193-208.
- Baba, Naohiko, Frank Packer, and Teppei Nagano. 2008. “The Spillover of Money Market Turbulence to FX Swap and Cross-Currency Swap Markets.” BIS Quarterly Review (March).
- Bank for International Settlements. 2019. “Reporting Guidelines for the BIS International Banking Statistics .” Bank for International Settlements, Basel.
- Basel Committee on Banking Supervision. 1996. “Amendment to the Capital Accord to Incorporate Market Risks.” Bank for International Settlements, Basel.
- Brauning, Falk and Victoria Ivashina. 2017. “Monetary Policy and Global Banking.” NBER Working Paper 23316.
- Bruno, Valentina, and Hyun Song Shin. 2015. “Cross-Border Banking and Global Liquidity.” Review of Economic Studies 82(2): 535-564.
- Bruno, Valentina and Hyun Shin. 2019. "Dollar exchange rate as a credit supply factor—evidence from firm-level exports." BIS Working Papers, No. 819
- Cerutti, Eugenio M., Stijn Claessens, and Patrick McGuire. 2012. “Systemic Risks in Global Banking: What Available Data Can Tell Us, and More Data are Needed?” NBER Working Paper 18531.
- Cerutti, Eugenio M., Maurice Obstfeld, and Haonan Zhou. 2019. “Covered Interest Parity Deviations—Macrofinancial Determinants.” IMF Working Paper 19/14, International Monetary Fund, Washington, DC.
- Du, Wenxin, Alexander Tepper, and Adrien Verdelhan. 2018. “Deviations from Covered Interest Rate Parity.” The Journal of Finance 73(3): 915–957.
- Gopinath, Gita and Jeremy Stein. 2018. “Banking, Trade, and the Making of a Dominant Currency,” National Bureau of Economic Research, Cambridge, Massachussetts.
- Gustavo Adler, Camila Casas, Luis Cubeddu, Gita Gopinath, Nan Li, Sergii Meleshchuk, Carolina Osorio Buitron, Damien Puy, and Yannick Timmer, forthcoming, “Dominant Currencies and External Adjustment,” IMF Staff Discussion Note.
- Goldberg, Linda S., Craig Kennedy, and Jason Miu. 2011. “Central Bank Dollar Swap Lines and Overseas Dollar Funding Costs.” Economic Policy Review 17 (1).
- Hofstetter, Marc, José Ignacio López, and Miguel Urrutia. 2018. “Limits to Foreign Exchange Net Open Positions and Capital Requirements in Emerging Economies.” Documentos CEDE 2018–10, Universidad de los Andes, Bogotá.
- Iida, Tomoyuki, Takeshi Kimura, and Nao Sudo. 2018. “Deviations from Covered Interest Rate Parity and the Dollar Funding of Global Banks.” International Journal of Central Banking 14 (4): 275–325.
- Ivashina, Victoria, David S. Scharfstein, and Jeremy C. Stein. 2015. “Dollar Funding and the Lending Behavior of Global Banks.” Quarterly Journal of Economics 130 (3): 1241–81.
- McGuire, Patrick and Goetz von Peter. 2009. “US Dollar Shortage in Global Banking,” BIS Quarterly Review (March).
- McGuire, Patrick, and Goetz von Peter. 2012. “The Dollar Shortage in Global Banking and the International Policy Response.” International Finance 15 (2): 155–78.
- Nakamura, Emi, and Jón Steinsson. 2018. "High-frequency identification of monetary non-neutrality: the information effect." The Quarterly Journal of Economics 133, no. 3: 1283-1330.
- Shin, H. 2012. “Global Banking Glut and Loan Risk Premium. IMF Economic Review 60(2) 155-192.
- Sushko, Vladyslav, Claudio E. V. Borio, Robert N. McCauley, and Patrick McGuire. 2016. “The Failure of Covered Interest Parity: FX Hedging Demand and Costly Balance Sheets.” BIS Working Paper 59, Bank for International Settlements, Basel.

### Figures and notes
- Figure 1. International Position vs. Foreign Position
  - Source: Cerutti, Claessens, and McGuire (2012).
  - Notes: In our analysis A corresponds to U.S. dollar cross-border claims; B refers to U.S. dollar local positions outside the U.S. (hence U.S. dollar is the foreign currency of the jurisdiction); C refers to the local positions in the U.S (i.e. non-U.S. banks’ branches and subsidiaries in the U.S.) - due to data limitations, this component, (C) is sourced from FFIEC002 filings for non-US banks’ U.S. branches and from call reports (i.e. FFIEC 031/041) for non-US banks’ U.S. subsidiaries when the balance sheets are constructed.
- Figure 2. Trends in US Dollar Activities of Non-US Banks
  - Panel titles:
    - 1. Non-US Banks’ US Dollar–Denominated Claims (Trillions of US dollars)
    - 2. Non-US Banks' Relative Share of US Branches and Subsidiaries in Total US Dollar–Denominated Claims, Latest Available (Percent of foreign position; bubble size = total assets)
    - 3. Non-US Banks’ US Dollar–Denominated Claims (Trillions of US dollars; excluding intragroup claims)
    - 4. Share of US Dollar–Denominated Claims of Non-US Banks ( Percent of total banking system assets)
  - Sources: Bank for International Settlements, locational banking statistics (nationality basis); Federal Financial Institutions Examination Council; S&P Global, Market Intelligence; and IMF staff calculations.
  - Notes: Foreign position consists of international position as defined by the Bank for International Settlements plus the positions in US branches and subsidiaries. The measure of US dollar-denominated claims, based on BIS data and represented in all four panels, may be larger in some cases than the trust account adjusted measure (see Saito, Hiyama, and Shiotani, 2018). In panel 1, some economies enter the sample after 2000. For example, China and Russia enter in 2015 and therefore account for a discrete jump in the series. Diagonal lines in panels 2 and 4 are 45-degree lines. Data labels in panel 4 use International Organization for Standardization (ISO) country codes.
  - Chart time labels shown: 2000:Q1, 01:Q1, 02:Q1, 03:Q1, 04:Q1, 05:Q1, 06:Q1, 07:Q1, 08:Q1, 09:Q1, 10:Q1, 11:Q1, 12:Q1, 13:Q1, 14:Q1, 15:Q1, 16:Q1, 17:Q1, 18:Q1, and markers "2018:Q1 or latest", "2010:Q4".
- Figure 3. US Dollar Funding Fragility of Non-US Banks
  - Panel titles:
    - 1. Non-US Banks’ US Dollar Cross-Currency Funding (Percent left scale; trillions of US dollars, right scale)
    - 2. Decomposition of Non-US Banks’ US Dollar Liquidity Ratio (Percent; computed at the aggregate level)
    - 3. Non-US Banks’ Liquidity Ratios: US Dollar Compared with All Currencies (Percent)
    - 4. Decomposition of Non-US Banks’ US Dollar Stable Funding Ratio (Percent, latest available)
  - Sources: Bank for International Settlements, locational banking statistics (nationality basis); Federal Financial Institutions Examination Council; S&P Global, Market Intelligence; and IMF staff calculations.
  - Notes: All panels correspond to the international position plus US branches of the non-US banks. Latest available calculations were as of 2018:Q1 at the time the analysis was conducted. Panel 1 shows the difference between US dollar assets and liabilities (both in trillions of dollars and as a percentage of US dollar assets) for a balanced sample of 13 economies. Panels 2 through 4 are based on a set of 14 economies that permit the calculation of the liquidity and stable funding ratios. Panels 2 and 4 are computed using the sample-wide aggregate values; the changes are in percentage points.
  - Chart time labels shown: 2000:Q1 through 2018:Q1 and scales labeled "Trillions of US dollars", "Percent", with series labels including "Cross-currency funding ratio (left scale)" and "Non-US Banks' cross-currency funding gap (right scale)".
- Figure 4. Interaction Effects of the Cross-Currency Funding Ratio and Drivers of the Cross-Currency Basis
  - Panel title:
    - 1. Three-Month Cross-Currency Basis (CCB) (Basis points, monthly average)

*Source: wpiea2020113-print-pdf - References*

### 2. Interaction Effects of the Cross-Currency Funding

### Interaction Effects of the Cross-Currency Funding Ratio and Drivers of the Cross-Currency Basis (Standardized coefficients)

### Panel summary (visual/chart notes)
- Panel 1: Monthly averages of the three-month Libor cross-currency basis, measured in selected currencies (sample period from 1/1/2000 to 03/01/2018).
- Panel 2: Aggregate impact of basis determinants with interaction at “low” (25th percentile) and “high” (75th percentile) CCFR. Panel 2 sample: AUD, CAD, CHF, EUR, GBP, HKD, JPY, MYR, SEK.
- Colored bars denote significance levels at 10 percent or lower.

### Key drivers and baseline associations (Table 1 — The Drivers of the Cross-Currency Basis)
- Dependent variables considered: 3m CCB (LIBOR), 1m CCB (LIBOR), 1y CCB (LIBOR), 1m CCB (OIS), 3m CCB (OIS), 1y CCB (OIS).
- Cross-Currency Funding Ratio (CCFR)
  - 3m CCB (LIBOR): -0.096***
  - 1m CCB (LIBOR): -0.252***
  - 1y CCB (LIBOR): 0.015
  - 1m CCB (OIS): -0.307***
  - 3m CCB (OIS): -0.126***
  - 1y CCB (OIS): 0.017
- LIBOR-OIS spread
  - 3m CCB (LIBOR): -0.131***
  - 1m CCB (LIBOR): -0.149***
  - 1y CCB (LIBOR): -0.058***
  - 1m CCB (OIS): -0.046
  - 3m CCB (OIS): 0.036
  - 1y CCB (OIS): -0.034***
- Term spread differential
  - 3m CCB (LIBOR): 0.195***
  - 1m CCB (LIBOR): 0.196***
  - 1y CCB (LIBOR): 0.014***
  - 1m CCB (OIS): 0.257***
  - 3m CCB (OIS): 0.246***
  - 1y CCB (OIS): 0.019***
- FX implied volatility
  - 3m CCB (LIBOR): -0.169***
  - 1m CCB (LIBOR): -0.098***
  - 1y CCB (LIBOR): -0.062***
  - 1m CCB (OIS): -0.120***
  - 3m CCB (OIS): -0.161***
  - 1y CCB (OIS): -0.069***
- Economy i default probability
  - 3m CCB (LIBOR): -0.080*
  - 1m CCB (LIBOR): -0.064
  - 1y CCB (LIBOR): -0.073***
  - 1m CCB (OIS): 0.031
  - 3m CCB (OIS): -0.010
  - 1y CCB (OIS): -0.055***
- US dollar index
  - 3m CCB (LIBOR): -0.028**
  - 1m CCB (LIBOR): -0.050***
  - 1y CCB (LIBOR): -0.009**
  - 1m CCB (OIS): -0.030*
  - 3m CCB (OIS): -0.025*
  - 1y CCB (OIS): -0.050***
- VIX
  - 3m CCB (LIBOR): 0.100***
  - 1m CCB (LIBOR): 0.102***
  - 1y CCB (LIBOR): 0.029***
  - 1m CCB (OIS): 0.047*
  - 3m CCB (OIS): -0.018
  - 1y CCB (OIS): -0.037***
- Observations and fit (selected)
  - 3m CCB (LIBOR): Observations 1,776; R-squared 0.882
  - 1m CCB (LIBOR): Observations 1,621; R-squared 0.857
  - 1y CCB (LIBOR): Observations 1,418; R-squared 0.633

Notes: Coefficients are standardized (impact of a one standard deviation increase). Robust standard errors in parentheses. Significance: *** p<0.01, ** p<0.05, * p<0.1.

### Interaction effects of CCFR with drivers (Table 2)
- Interaction coefficients (selected, standardized):
  - LIBOR-OIS spread x CCFR
    - 3m CCB (LIBOR): -0.124***
    - 1m CCB (LIBOR): -0.072*
    - 1y CCB (LIBOR): -0.056***
  - Term spread differential x CCFR
    - 3m CCB (LIBOR): 0.174***
    - 1m CCB (LIBOR): 0.252***
    - 1m CCB (OIS): 0.280***
  - Bid-ask spread x CCFR
    - 3m CCB (LIBOR): -0.040**
    - 1m CCB (LIBOR): -0.105*
    - 1y CCB (LIBOR): 0.021**
  - FX implied volatility x CCFR
    - 3m CCB (LIBOR): -0.137***
    - 1y CCB (LIBOR): -0.146***
    - 1y CCB (OIS): -0.179***
  - Economy i default probability x CCFR
    - 3m CCB (LIBOR): -0.098*
    - 1y CCB (LIBOR): -0.070***
  - US dollar index x CCFR
    - 3m CCB (LIBOR): -0.213*
    - 1y CCB (LIBOR): -0.318***
    - 1y CCB (OIS): -0.327***
  - VIX x CCFR
    - 3m CCB (LIBOR): 0.308***
    - 1m CCB (LIBOR): 0.284***
    - 1y CCB (LIBOR): 0.107***
    - 1m CCB (OIS): 0.217***
- Observations and fit (selected)
  - 3m CCB (LIBOR): Observations 1,820; R-squared 0.888
  - 1m CCB (LIBOR): Observations 1,827; R-squared 0.866
  - 1y CCB (LIBOR): Observations 1,813; R-squared 0.626

Notes: Interaction coefficients standardized. Estimates of levels not shown. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1.

### Effects of drivers at low vs high CCFR (Table 3)
- Comparison of the effects (standardized) for Low CCFR and High CCFR:
  - Term spread differential: Low CCFR 0.275; High CCFR 0.510
  - US dollar index: Low CCFR 0.086; High CCFR -0.010
  - VIX: Low CCFR 0.069; High CCFR 0.515
  - LIBOR-OIS spread: Low CCFR -0.016; High CCFR -0.217
  - Economy i default probability: Low CCFR -0.156; High CCFR -0.303
  - Bid-ask spread: Low CCFR -0.106; High CCFR -0.145
  - FX implied volatility: Low CCFR -0.224; High CCFR -0.411

### Changes in CCFR–basis relationship around US financial regulations (Table 4)
- CCFR (Pre-reform)
  - 3m CCB (LIBOR): -0.059***
  - 1m CCB (LIBOR): -0.070***
  - 1y CCB (LIBOR): -0.138***
- CCFR x 1{Stress VaR} (Stressed VaR, 2013): 0.027*** (3m CCB)
- CCFR x 1{Supplementary leverage ratio} (2014): 0.021*** (1y CCB)
- CCFR x 1{Liquidity coverage ratio} (2015): -0.029*** (3m CCB)
- CCFR x 1{MMMF reform} (2016): -0.042*** (3m CCB)
- Observations: 420 (all columns). R-squared range: 0.610 to 0.791.
Notes: Pre-Reform corresponds to 2012. Model at monthly frequency; controls include standard CCB drivers. Coefficients standardized. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1.

### US dollar funding costs and home economy financial stress (Tables 5–7)
- Table 5: Association of ΔlogPD with ΔFunding Cost (ΔDollar Funding cost = negative of CCB)
  - Column (1): ∆Funding Cost 0.142** (observations 1,080; R-squared 0.102)
  - Column (2): ∆Funding Cost 0.133**; ∆Funding Cost^2 0.018* (R-squared 0.108)
  - Column (3): ∆Funding Cost X period dummies: coefficients include 0.213**, 0.302***, and -0.006 for respective period interactions (R-squared 0.115)
  - Notes: ΔlogPD and ΔDollar Funding cost standardized. Controls: macro and bank controls; moving averages t-4 to t-1.
- Table 6: Instrumental variable robustness (US monetary policy shocks)
  - ∆Funding Cost (instrumented with Fed funds rate shock): 4.525 (Std. error 2.631) — not starred
  - ∆Funding Cost (instrumented with Policy News Shock): 7.128** (Std. error 2.815)
  - Observations 724; R-squared 0.220
  - Notes: Two-stage IV using Federal Funds rate shock and policy news shock following Nakamura and Steinsson (2018).
- Table 7: Amplification by USD funding vulnerability measures
  - Selected coefficients:
    - Column (2) (CCFR specification): ∆Funding Cost 0.122**; ∆Funding Cost X CCFR, lag annual average 0.071 (observations 1,080; R-squared 0.108)
    - Column (4) (LCR specification): ∆Funding Cost -0.142; ∆Funding Cost X Quintile LCR, lag annual average 0.090** (observations 587; R-squared 0.118)
    - Column (5) (SFR specification): ∆Funding Cost -0.148; ∆Funding Cost X Quintile SFR, lag annual average 0.099*** (observations 587; R-squared 0.129)
  - Notes: ΔlogPD, ΔDollar Funding cost, US dollar asset share and CCFR standardized; LCR and SFR expressed as quintiles (higher = lower liquidity/stable funding). Controls as in prior tables.

### Swap lines, domestic bank health, and international reserves (Tables 8–11)
- Table 8: Effect of Swap Lines with the US Federal Reserve on ΔlogPD
  - ∆Funding Cost X 0.(SWAP Announcement Dummy): 0.174*** (Column 1), 0.154*** (Column 2)
  - ∆Funding Cost X 0.(SWAP Dummy): 0.171*** (Column 3), 0.152*** (Column 4)
  - SWAP Announcement Dummy and SWAP Dummy coefficients not significant in presented columns.
  - Observations 1,080; R-squared 0.104–0.108.
- Table 9: Mitigating effect of domestic bank health on ΔlogPD
  - Bank fundamentals (lag annual averages)
    - Bank Capital/Asset: coefficients ~0.048–0.080 (not significant)
    - Bank Cash/Asset: coefficients ~-0.060 (not significant)
    - Bank ROA: 0.239***, 0.229***, 0.230*** (highly significant)
  - ∆Funding Cost: ranges 0.406***, 0.404***, 0.325***, 0.386***, 0.397***, 0.304*** across columns (all significant)
  - Interaction terms (∆Funding Cost X bank fundamentals)
    - ∆Funding Cost X Bank Capital/Asset: -0.082***; -0.080*** (mitigating)
    - ∆Funding Cost X Bank Cash/Asset: -0.131***; -0.144*** (mitigating)
    - ∆Funding Cost X Bank ROA: -0.083***; -0.082*** (mitigating)
  - Observations 1,080; R-squared 0.112–0.118.
- Table 10: International reserves and mitigation of amplification (LCR/SFR specifications)
  - ∆Funding Cost: -0.917** (Column 1 LCR), -0.488** (Column 2 SFR), similar magnitudes in Columns (3)-(4).
  - International Reserve/GDP, lag annual average: -0.059; -0.201** (SFR specification)
  - ∆Funding Cost X International Reserve/GDP, lag annual average: 0.599***; 0.278* (Columns 1–2)
  - ∆Funding Cost X Quintile USD Fragility Ratio, lag annual average: 0.311***; 0.207*** (Columns 1–2)
  - Triple interaction ∆Funding Cost X International Reserve/GDP X USD Fragility Ratio: -0.175***; -0.088** (Columns 1–2)
  - Observations 587; R-squared 0.133–0.139.
  - Notes: LCR and SFR expressed as quintiles; all controls as prior.

### Cross-border US dollar lending and substitution (Tables 11–14)
- Table 11: Impact of US dollar funding shocks on cross-border US dollar lending (bilateral lending i→j)
  - Baseline (Column 1): ∆Funding Cost -5.311*** (Std. error 1.683)
  - EM Lenders (Column 2): ∆Funding Cost 6.890 (not significant)
  - EM Recipients (Column 3): ∆Funding Cost -3.073 (not significant)
  - Quadratic test (Column 4): ∆Funding Cost -3.920***; ∆Funding Cost^2 -0.271*** (non-linearity)
  - Initial Cost interaction (Column 5): ∆Funding Cost -2.725**; ∆Funding Cost x Initial Cost -0.436**
  - Observations 23,154; R-squared 0.953.
  - Notes: Funding cost shock in 50 bps per annum; log dependent variable multiplied by 100 (impact in percentage points).
- Table 12: Substitution across lenders and between cross-border and domestic USD lenders
  - ∆Funding cost of lender i: 0.113*** (Column 1); 0.132*** (Column 2)
  - ∆Funding cost of other lenders (-i): 3.156* (Column 1); 7.826*** (Column 2)
  - ∆Funding cost of all lenders: -2.800***; -3.097*** in some specifications
  - Observations 22,802 (Columns 1–2); smaller samples for other columns.
  - Notes: Dependent variables vary by column; funding cost in 50 bps; effects in percentage points.
- Table 13: Amplification of funding vulnerability on transmission to cross-border USD lending
  - Column (1) CCFR specification:
    - ∆Funding Cost 0.617
    - ∆Funding Cost x CCFR -0.758*
    - Observations 23,840; R-squared 0.916
  - Column (2) LCR specification:
    - ∆Funding Cost -3.882***; ∆Funding Cost x LCR 1.195**
    - Observations 17,216; R-squared 0.931
  - Column (3) SFR specification:
    - ∆Funding Cost -2.169***; ∆Funding Cost x SFR 0.874**
    - Observations 17,216; R-squared 0.931
  - Notes: Vulnerability measures in quintiles; funding cost shock in 50 bps; dependent variable scaled to percentage points.
- Table 14: Mitigating effects on transmission to cross-border USD lending
  - Selected coefficients (Columns and observations vary):
    - ΔFunding Cost: ranges from -6.615*** to -1.603 across columns
    - ΔFunding Cost x Bank Liquidity Ratio: 1.951***; 2.128*** (mitigating/modulating)
    - ΔFunding Cost x Bank Return on Assets (ROA): 0.934**; 0.779** 
    - ΔFunding Cost x Swap line: 7.544**; 9.404*** (positive interaction)
    - ΔFunding Cost x Int'l Reserves: 0.313***; 0.292*** (Columns with reserves)
    - ∆Funding Cost x CCFR: coefficient examples include 0.039; -1.022**; -0.553; 0.039 (heterogeneous)
  - Observations range 20,356 to 23,154 across columns; R-squared 0.922–0.953.
  - Notes: Dependent variable logged and multiplied by 100; robustness with clustered standard errors at lender and recipient levels.

### Recipient financial stress and lending (Table 15)
- Association of recipient economy ΔlogPD with weighted average ∆Funding Cost (lenders)
  - All recipients (Column 1): ∆Funding Cost (Weighted Average) 0.095* (observations 874; R-squared 0.031)
  - Top 10 recipients (Column 2): 0.146* (observations 230; R-squared 0.037)
  - Rest recipients (Column 3): 0.091 (observations 644; R-squared 0.029)
  - Notes: ΔlogPD and ΔDollar Funding cost standardized; controls include macro and bank variables averaged t-4 to t-1.

### Appendix reference
- Figure A 1: Testing for endogenous relationship between CCFR and CCB (figure referenced in appendix).

*Source: IMF staff calculations (content unit: 2. Interaction Effects of the Cross-Currency Funding Ratio and Drivers of the Cross-Currency Basis, wpiea2020113-print-pdf).*

### 1. IRF: CCFR →3m CCB

### 1. IRF: CCFR →3m CCB

### Impulse Response Functions (IRFs) — summary of figures
- The figure reports the impulse response functions (IRF) of CCB to one standard deviation shock of selected CCB drivers:
  - 1. IRF: CCFR →3m CCB (percent)
  - 2. IRF: Term Spread Differential →3m CCB (percent)
  - 3. IRF: Home Country EDF →3m CCB (percent)
  - 4. IRF: USD Index →3m CCB (percent)
- Estimation details:
  - Coefficients are estimated using a panel VAR of order three with the following Cholesky order: term spread differential, USD dollar index, VIX, CCFR, FX implied volatility, BID-ASK spread, Economy i default probability, Libor-OIS spread and three-month CCB.
  - Confidence level is 95 percent.
- Visual axes and tick markers preserved as presented:
  - For panels with horizontal axis ticks: 0 2 4 6 8 10 12 14 16 18
  - Vertical axis numeric ranges shown in the figure include, for different panels, values such as:
    - -0.10, -0.08, -0.06, -0.04, -0.02, 0.00, 0.02
    - 0.00, 0.02, 0.04, 0.06, 0.08, 0.10, 0.12
    - -0.20, -0.15, -0.10, -0.05, 0.00, 0.05, 0.10
    - -0.04, -0.02, 0.00, 0.02, 0.04, 0.06, 0.08

### Notes accompanying the IRF figure
- Source: IMF staff estimates.
- The figure shows responses of three-month CCB to one-standard-deviation shocks in the listed drivers.
- Estimation: panel VAR of order three with Cholesky ordering as listed above.
- Confidence intervals shown at the 95 percent level.

### Table A.2 — Amplification Effect of US Dollar Funding Vulnerability on the Relationship between Funding Costs and Home Economy Financial Stress (Euro Area)
- Dependent variable: ∆logPD
- Sample: Euro area countries
- The table shows that the impact of increases in US dollar funding costs on changes in the log probability of default (ΔlogPD) of the home economy banking sector is amplified by higher USD funding fragility.
- Specification matches the full sample (Table 7) with alternative measures of USD funding vulnerability across columns.

Key coefficient estimates and standard errors (standard errors in parentheses):
- Column (1) — USD Asset Share
  - ∆Funding Cost: -0.028 (0.146)
  - USD Assets/Total Assets, lag annual average: 0.010 (0.505)
  - ∆Funding Cost X (USD Assets/Total Assets, lag annual average): 0.097 (0.110)
  - Observations: 363
  - R-squared: 0.134
- Column (2) — CCFR
  - ∆Funding Cost: 0.047 (0.044)
  - CCFR, lag annual average: -0.484 (0.247)
  - ∆Funding Cost X (CCFR, lag annual average): 0.113* (0.052)
  - Observations: 363
  - R-squared: 0.152
- Column (3) — Sign dummy of CCFR
  - ∆Funding Cost: -0.082 (0.104)
  - 1.[CCFR, lag annual average>=0]: 0.614* (0.243)
  - ∆Funding Cost X 0.[CCFR, lag annual average>=0]: 0.216 (0.114)
  - ∆Funding Cost X 1.[CCFR, lag annual average>=0]: -0.275 (0.206)
  - CCFR, lag annual average X 0.[CCFR, lag annual average>=0]: -0.651 (0.325)
  - CCFR, lag annual average X 1.[CCFR, lag annual average>=0]: -1.038* (0.420)
  - ∆Funding Cost X CCFR, lag annual average X 0.[CCFR, lag annual average>=0]: 0.350 (0.174)
  - ∆Funding Cost X CCFR, lag annual average X 1.[CCFR, lag annual average>=0]: 0.494* (0.235)
  - Observations: 363
  - R-squared: 0.180
- Column (4) — USD Liquidity Ratio (LCR)
  - ∆Funding Cost: -0.002 (0.119)
  - Quintile LCR, lag annual average: 0.015 (0.059)
  - ∆Funding Cost X Quintile LCR, lag annual average: 0.057* (0.027)
  - Observations: 261
  - R-squared: 0.169
- Column (5) — USD Stable Funding Ratio (SFR)
  - ∆Funding Cost: (not separately listed beyond column header) -0.002 reported under column (4)/(5) layout
  - Quintile SFR, lag annual average: 0.010 (0.047)
  - ∆Funding Cost X Quintile SFR, lag annual average: 0.027 (0.030)
  - Observations: 282
  - R-squared: 0.159

Panel-level controls and clustering:
- Country fixed effects: Yes (all columns)
- Macro controls: Yes (all columns)
- Bank controls: Yes (all columns)
- Standard errors are clustered at the economy level in all regressions.

Statistical significance notation:
- *** p<0.01, ** p<0.05, * p<0.1

### Interpretation of main findings
- Higher CCFR (US dollar funding gap/US dollar assets) amplifies the effect of increases in US dollar funding costs on banking-sector ΔlogPD:
  - In Column (2), the interaction term ∆Funding Cost X (CCFR, lag annual average) = 0.113* (0.052) indicates a positive and marginally significant amplification.
- Nonlinear/sign-dependent effects captured in Column (3) suggest the sign and magnitude of CCFR matter:
  - 1.[CCFR, lag annual average>=0] = 0.614* (0.243)
  - CCFR, lag annual average X 1.[CCFR, lag annual average>=0] = -1.038* (0.420)
  - ∆Funding Cost X CCFR, lag annual average X 1.[CCFR, lag annual average>=0] = 0.494* (0.235)
- Liquidity and stable-funding quintiles show weaker amplification effects:
  - ∆Funding Cost X Quintile LCR, lag annual average = 0.057* (0.027)
  - ∆Funding Cost X Quintile SFR, lag annual average = 0.027 (0.030) (not significant at conventional levels)

### Observations on model fit and sample size
- Observations by column: 363, 363, 363, 261, 282
- R-squared by column: 0.134, 0.152, 0.180, 0.169, 0.159

*Source: IMF staff estimates. Notes: Focusing on euro area countries, this table shows that the impact of increases in US dollar funding costs on changes in the log probability of default (ΔlogPD) of the home economy banking sector is amplified by higher USD funding fragility. Specification is the same used for the full sample (Table 7). Results in Column (1) correspond to specification with US dollar asset share. Results in Column (2) correspond to the specifications with US dollar funding gap/US dollar assets (CCFR). Results in Column (3) correspond to the specifications with sign dummy of US dollar funding gap/US dollar assets (CCFR). Results in Column (4) correspond to the specification with USD liquidity ratio (LCR). Results in Column (5) correspond to the specification with USD stable funding ratio (SFR). Standard errors are clustered at the economy level in all regressions. Standard errors are clustered at the economy level in all regressions. *** p<0.01, ** p<0.05, * p<0.1.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020113-print-pdf.pdf_