## wpiea2025065-print-pdf

## Source details

**Canonical URL:** [wpiea2025065-print-pdf](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025065-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025065-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025065-print-pdf.pdf.json)

---

### Defining UIP deviations
- UIP deviations for a currency pair j/$ are defined as residuals, λ_{j/$,k}_t, from the classic UIP condition:
  - λ_{j/$,k}_t = ln(1+i_{j,k}_t) − ln(1+i_{$,k}_t) − (ln(E(S_{j/$,t+k})) − ln(S_{j/$,t})), (equation (1))
  - Nominal spot exchange rate, S_{j/$,t}, and expected exchange rate at horizon k, E(S_{j/$,t+k}), are expressed in local currency j units per US dollar (an increase represents a US dollar appreciation).
  - If λ_{j/$,k}_t = 0, UIP holds; λ_{j/$,k}_t > 0 implies positive excess returns for the foreign currency over the US dollar.
- Data and measurement choices:
  - Focus: major advanced economies included in the Nominal Advanced Foreign Economies U.S. Dollar Index.
  - Forecasts: Consensus Economics monthly surveys of exchange rate forecasts covering 1999Q1-2023Q4.
  - Forecast horizon: k set at two years.
  - Risk-free rates: 2-year government bond yields.
  - Frequency: Monthly UIP deviations aggregated into quarterly averages.

### Constructing the aggregate index
- Aggregate weighted UIP deviations:
  - λ_{AE,t} = Σ_j ω_j λ_{j/$,t}, (equation (2))
  - USD index weights (ω_j) used for aggregation to compare directly to the USD index.
- Average weights for advanced economies in the index:
  - AUD 2.7, CAD 30.4, JPY 14.3, SWK 1.3, CHF 4.5, GBP 10.6, EUR 36 percent.

### Key empirical results on UIP deviations and the USD index
- Persistence and centrality:
  - The aggregate measure exhibits persistent deviations from zero.
  - Mean quarterly deviation: 0.005.
  - Absolute average quarterly mean deviation: 0.026.
- Strong positive correlations:
  - Level correlation between λ_{AE,t} and the USD index: 0.86.
  - Pre-GFC (1999-2007) correlation: 0.85.
  - Post-GFC (2010-2022) correlation: 0.82.
  - Correlation for quarterly changes: 0.80.
- Country-level correlations (λ_{j/$,t} with LC/USD, USD index, λ_{AE,t}):
  - GBP: 0.66, 0.45, 0.63
  - SWK: 0.80, 0.78, 0.91
  - CHF: 0.82, 0.66, 0.85
  - CAD: 0.81, 0.82, 0.74
  - JPY: 0.66, 0.28, 0.40
  - EUR: 0.84, 0.78, 0.95
  - AUD: 0.87, 0.82, 0.87
  - AE Index Average: 0.86, 1.00
  - Overall bilateral correlations lie in the 0.66-0.87 range; the Japanese yen is a notable exception with weaker correlations (driven by the pre-GFC sample).
- Aggregation and robustness notes:
  - Comparable results reported for 1-year and 3-months horizons (Annex A).
  - Annex A decomposition: most movements in λ_{AE,t} associated with expected exchange rate changes rather than interest rate differentials.

### Interpretation and plausible drivers
- Risk-premium channel:
  - When risk appetite falls, the US dollar appreciates, raising risk premia for other currencies; this generates positive excess returns for foreign currency j over the US dollar and the positive correlation between UIP deviations and the USD index.
- Global financial intermediaries:
  - Higher demand for US dollars may lead intermediaries to demand a higher expected return for supplying dollars, contributing to UIP deviations.
- Research caveat:
  - Ultimate sources of financial-market-driven US dollar fluctuations remain an active area of research and are beyond the present study.

### Comparison with other global financial indicators
- Correlations (quarterly levels, 1999:Q1–2022:Q4):
  - USD Index with UIP Deviations: 0.86
  - USD Index with Global Financial Cycle: −0.39
  - UIP Deviations with Global Financial Cycle: −0.25
  - USD Index with VIX: 0.16
  - UIP Deviations with VIX: 0.25
  - USD Index with Commodity Price Index: −0.55
  - UIP Deviations with Commodity Price Index: −0.56
- Interpretation:
  - The USD index–UIP deviations correlation is considerably stronger and more robust than correlations between the USD index or UIP deviations and commonly studied global financial condition measures.
  - The Global Financial Cycle correlation with the USD index weakens when the 2008-10 GFC period is excluded in levels; quarterly-difference correlations remain well below the USD–UIP relation.
  - VIX shows modest correlations (~0.2) with the USD index and UIP deviations over the full sample; VIX–USD index correlation is 0.73 only in the pre-GFC sub-period.
  - Commodity price index exhibits negative correlation with both the USD index and UIP deviations (−0.56).

### Appendix A — Horizon-specific and decomposition results (selected)
- 1-year horizon:
  - UIP deviations centered around zero, mean deviation of 0.002.
  - Mean absolute deviation: 0.020 (23% smaller than 2-year horizon 0.026).
  - Strong correlation with USD index; Japanese yen and British pound show weaker correlations.
  - Table A1 (1-Year correlations) entries preserved exactly as in source.
- 3-months horizon:
  - Deviations centered around zero at −0.005 and 56% smaller than 2-year horizon.
  - Aggregated measure correlation with USD index: 0.18.
  - Individual currency correlations at 3-months range from −0.38 to 0.57.
  - Table A2 (3-Months correlations) entries preserved exactly as in source.
- First-differences (2-year horizon):
  - Table A3 (∆(λ_j/$_t) correlations): GBP 0.66; SWK 0.76; CHF 0.80; CAD 0.85; JPY 0.68; EUR 0.77; AUD 0.84; AE Index Average – 0.80 1.00
- Decomposition (equation A.1):
  - λ_AE,k_t = Σ_j ω_j [ln(1+i_j,k_t) − ln(1+i_$,k_t)] + Σ_j ω_j [ln(S_j/$_t) − ln(E(S_j/$_t+k))]
  - Finding: bulk of UIP deviations related to expected exchange rate adjustment; interest rate differentials relatively subdued and less volatile for AEs over last 20 years.

### Appendix C — Instrument relevance and first-stage strength (selected)
- IV first-stage (changes in USD index on UIP deviations) controls from equation (3):
  - IV coefficients significant at 1 percent for both EMs and AEs.
  - Kleibergen-Paap Wald F-statistics examples:
    - Emerging Markets F-statistics: Min. 65.35, Median 159.66, Max. 254.70
    - Advanced Economies F-statistics: Min. 118.90, Median 164.68, Max. 194.81
  - Observations reported in table formatting: 779, 595, 959, 595, 959, 595.

---

### 4.2 Policy Regimes and Structural Characteristics in Emerging Markets — Overview
- Purpose: Analyze heterogeneous effects of US dollar appreciations across economies by policies and structural characteristics (monetary anchoring, exchange rate regimes, commodity dependence, trade openness, US dollar exposures).
- Method: State-dependent responses via sample splits or country-specific factor values; country factors captured as sample-period averages to avoid endogeneity.
- Identification challenges:
  - Many characteristics correlate strongly with EM vs AE split (notably monetary policy anchoring).
  - High collinearity among country characteristics (e.g., exchange rate regimes correlated with US dollar invoicing; commodity exporters correlated with higher USD liabilities, lower GVC participation, higher USD invoicing, less flexible exchange rates, and lower trade openness).
- Strategy:
  - Limit identification to heterogeneity within EMs.
  - Focus sequence: commodity dependence (slow moving) → monetary policy anchoring → conditional exchange rate regime analysis → other characteristics conditional on commodity dependence.

### Commodity Dependence — Key findings and exact quantified effects
- Key finding:
  - Commodity exporters exhibit larger negative spillovers owing to concurrent deteriorations in their terms of trade.
- Exact quantified effect:
  - A 1 percent US dollar appreciation decreases the terms of trade by 0.7 percent after three quarters.
- Dynamics and mechanisms:
  - Commodity importers’ terms of trade remain broadly unaffected; output decline is shallower; eventual REER depreciation buffers spillovers.
  - No evidence that REER depreciates disproportionately for commodity exporters to offset falling commodity prices (consistent with fear of floating).
  - No evidence of accommodative monetary policy in commodity exporters; they tighten monetary policy significantly from the 3rd quarter onward.
- Contrast with advanced economies:
  - AE commodity exporters show sizable negative terms-of-trade responses but do not experience more negative output impacts than AE commodity importers because AE commodity exporters allow significant REER depreciation and pursue more accommodative monetary policy.

### Monetary Policy Credibility — Key dynamics
- Key finding:
  - Monetary policy anchoring mitigates negative spillovers by facilitating accommodative policy responses.
- Exact dynamics:
  - EMs with more anchored inflation expectations exhibit a shallower initial decline in output (difference with less anchored EMs is statistically significant).
  - When inflation expectations anchored:
    - REER depreciates on impact.
    - Policy rate becomes more accommodative (policy rates decrease).
    - Countries decrease current account balances to smooth temporary output drops.
  - Less anchored EMs:
    - Policy rates increase.
    - REER appreciates on impact, contributing to larger negative spillovers.
- Note:
  - Controlling for commodity dependence does not affect these findings.

### Exchange Rate Flexibility — Exact dynamics and caveats
- Key finding:
  - Freely floating exchange rate regimes facilitate faster recoveries in output for EMs after a US dollar appreciation.
- Exact dynamics:
  - Freely floating EMs:
    - REER depreciates sizably on impact and remains depreciated into the medium term.
    - Avoid deflationary impact on CPI.
    - Credit recovers in medium term; private capital inflows outperform.
    - Current account decreases as output recovers.
  - Other (less flexible) regimes:
    - REER does not depreciate during first two years.
    - Persistent deflationary impulse on CPI.
    - Credit declines in medium term; current account tends to increase.
- Collinearity caveat:
  - Exchange rate regimes highly correlated with share of exports invoiced in US dollars (correlation of −0.9).
  - Findings for USD export invoicing mirror those for exchange rate flexibility:
    - Lower USD export invoicing share speeds up output recovery via REER depreciation.
    - High USD invoicing EMs tighten monetary policy, exhibit initial deflationary CPI impulse, and REER does not depreciate during initial 8 quarters.
  - Collinearity prevents clean separation of independent effects of exchange rate regime versus USD invoicing.

### Other characteristics (conditional on commodity dependence) — Selected findings and exact percentiles
- Trade openness:
  - High trade openness EMs: negative output spillovers shallower and limited to initial three quarters; REER depreciates on impact; sizable initial current account decrease smooths output impact.
  - Low trade openness EMs: protracted output fall extending to 10 quarters; significant monetary tightening and initial REER appreciation.
  - GVC participation highly correlated with trade openness (+0.9) and shows similar impulse responses.
- US dollar liability exposure:
  - More USD-denominated foreign liabilities tend to be associated with harder hits from USD appreciations.
  - Much of significant negative output response driven by commodity exporters that are also more USD-liability exposed.
  - Controlling for commodity dependence: EMs with higher USD liability exposure are only marginally more negatively affected in output than less exposed EMs.
  - Dynamics: Highly exposed EMs experience REER appreciation on impact and tighten monetary policy; low-exposure EMs depreciate REER in short run and decrease policy rates.
- Reserve assets:
  - Countries with higher stocks of reserve assets estimated to systematically intervene to buffer exchange rate movements.
  - No significant difference in short-run exchange rate or GDP responses to the shock based on reserve asset levels.
- Selected percentiles and country averages (Table B7):
  - Net Commodity Exports: Min -0.06, Max 0.17, Mean 0.03, 10th Pct -0.05, 25th Pct -0.03, Median 0.01, 75th Pct 0.07, 90th Pct 0.17
  - MP Credibility: Min -1.38, Max 0.33, Mean -0.13, 10th Pct -0.95, 25th Pct -0.18, Median 0.03, 75th Pct 0.11, 90th Pct 0.20
  - Trade Openness: Min 26.39, Max 157.57, Mean 65.35, 10th Pct 28.21, 25th Pct 43.10, Median 52.41, 75th Pct 72.21, 90th Pct 150.80
  - USD Export Invoicing: Min 17.21, Max 98.76, Mean 70.85, 10th Pct 25.85, 25th Pct 47.02, Median 81.85, 75th Pct 94.43, 90th Pct 97.12
  - GVC Participation: Min 0.09, Max 0.45, Mean 0.20, 10th Pct 0.10, 25th Pct 0.12, Median 0.18, 75th Pct 0.26, 90th Pct 0.36
  - USD Liab. to Total Liab.: Min 0.09, Max 0.51, Mean 0.30, 10th Pct 0.12, 25th Pct 0.18, Median 0.33, 75th Pct 0.36, 90th Pct 0.45
  - FX Reserves: Min 0.05, Max 0.37, Mean 0.18, 10th Pct 0.09, 25th Pct 0.11, Median 0.16, 75th Pct 0.23, 90th Pct 0.36

### Robustness
- Sensitivity checks preserved baseline results:
  - Adding various monetary policy shock series (one at a time) does not change baseline results.
  - Controlling for oil supply shocks from Baumeister and Hamilton (2019) yields similar results.
  - Dropping fixed exchange rate regimes does not drive EM results.
  - Restricting sample to pre-COVID19 period (sample ends in 2019Q4) yields stronger rebound in output one year after shock but preserves output response differences between EMs and AEs.
  - Removing global controls increases magnitude of output declines for both AEs and EMs, but AEs still experience shallower, shorter-lived contraction.
  - Excluding all AEs included in the US dollar trade weighted index does not materially change point estimates; error bands increase.

### Conclusion and policy implications
- Aggregate finding:
  - Negative spillovers from US dollar appreciations are more pronounced in emerging markets, with larger and longer-lived declines in output than in advanced economies.
  - REER depreciation facilitates adjustment in advanced economies; in EMs, REER does not adjust on impact and depreciates only gradually (consistent with fear of floating).
  - Financial channels (reduced capital inflows, decline in domestic credit) contribute to adverse spillovers in EMs.
- Role of commodity exposure:
  - Commodity export exposure magnifies spillovers because US dollar appreciations are historically associated with deteriorating commodity prices, producing terms-of-trade declines not offset by REER depreciation in EM commodity exporters.
- Policy recommendations for EMs:
  - Anchor inflation expectations to mitigate negative spillovers through accommodative monetary responses, REER depreciation, and policy rate decreases.
  - Adopt and maintain exchange rate flexibility to speed up economic recovery following US dollar appreciations.
- Broader research implication:
  - Understanding multilateral policies affecting the global dollar cycle requires deeper analysis of uncovered UIP deviations—linked to market-wide risk appetite and risk premia demanded by global financial intermediaries—which reflect intermediary frictions and spillovers from financial regulation.

### Chapter 8 — Model simulations (FSGM) — setup and main model outcomes (selected)
- Model and calibration:
  - Uses FSGM (Andrle et al., 2015), semistructural multiregion general equilibrium model; analysis uses G20MOD module covering every G20 economy.
  - Key modeling features include inflation-forecast-based monetary rules, UIP deviations modeled as risk premiums, commodities priced in US dollars, and producer pricing.
- Simulation shock:
  - Global persistent 1 percentage point shock to the sovereign premium (applied excluding the United States).
- Main model outcomes:
  - Sovereign premium shock generates US dollar appreciation via increased demand for US dollars as risk-free returns on foreign bonds fall.
  - Financing costs increase, reducing domestic consumption and investment; output falls outside the US, larger in emerging markets due to limited exchange rate flexibility.
  - Commodity prices decline: a 1 percent appreciation in the US dollar associated with a 2.3 percent decline in commodity prices at a one-year horizon in the simulation.
    - US dollar pricing channel accounts for about 10 percent of overall fall in commodity price after one year in model decomposition.
  - Global trade openness falls as investment drops and imports decline.
  - Current account effects:
    - Commodity importers: terms-of-trade improvement raises real income; current account increases, more so for EM commodity importers because of larger fall in investment.
    - Commodity exporters: opposing forces broadly offset, leaving current account unchanged in the simulation.
- Model caveats:
  - Omits detailed balance sheet mismatches and intermediary frictions; UIP deviations serve as proxy for risk premiums rather than explicitly modeling intermediary rents.

*Source: IMF staff analysis (sections 2.1 and 4.2; Chapter 8 material) from wpiea2025065-print-pdf*

### 2.1    Constructing a Measure of UIP Deviations

### 2.1    Constructing a Measure of UIP Deviations

### Defining UIP deviations
- UIP deviations for a currency pair j/$ are defined as residuals, λ_{j/$,k}_t, from the classic UIP condition:
  - λ_{j/$,k}_t = ln(1+i_{j,k}_t) − ln(1+i_{$,k}_t) − (ln(E(S_{j/$,t+k})) − ln(S_{j/$,t})), (equation (1))
  - Nominal spot exchange rate, S_{j/$,t}, and expected exchange rate at horizon k, E(S_{j/$,t+k}), are expressed in local currency j units per US dollar (an increase represents a US dollar appreciation).
  - If λ_{j/$,k}_t = 0, UIP holds; λ_{j/$,k}_t > 0 implies positive excess returns for the foreign currency over the US dollar.
- Data and measurement choices:
  - Focus on major advanced economies included in the Nominal Advanced Foreign Economies U.S. Dollar Index.
  - Use Consensus Economics monthly surveys of exchange rate forecasts covering 1999Q1-2023Q4.
  - Forecast horizon k is set at two years to limit subjectivity in exchange rate forecasts.
  - Risk-free interest rates are measured with 2-year government bond yields.
  - Monthly UIP deviations are aggregated into quarterly averages to match macro data.

### Constructing the aggregate index
- Aggregate weighted UIP deviations are constructed as:
  - λ_{AE,t} = Σ_j ω_j λ_{j/$,t}, (equation (2))
  - USD index weights (ω_j) are used for aggregation so the aggregated UIP measure can be compared directly to the USD index.
- Average weights for advanced economies in the index are: AUD 2.7, CAD 30.4, JPY 14.3, SWK 1.3, CHF 4.5, GBP 10.6, EUR 36 percent.

### Key empirical results on UIP deviations and the USD index
- Persistence and centrality:
  - The aggregate measure of UIP exhibits persistent deviations from zero.
  - Mean quarterly deviation: 0.005.
  - Absolute average quarterly mean deviation: 0.026.
- Strong positive correlation with the USD index:
  - Level correlation between λ_{AE,t} and the USD index: 0.86.
  - Pre-GFC (1999-2007) correlation: 0.85.
  - Post-GFC (2010-2022) correlation: 0.82.
  - Correlation for quarterly changes: 0.80.
- Country-level correlations (λ_{j/$,t} with LC/USD, USD index, λ_{AE,t}):
  - GBP: 0.66, 0.45, 0.63
  - SWK: 0.80, 0.78, 0.91
  - CHF: 0.82, 0.66, 0.85
  - CAD: 0.81, 0.82, 0.74
  - JPY: 0.66, 0.28, 0.40
  - EUR: 0.84, 0.78, 0.95
  - AUD: 0.87, 0.82, 0.87
  - AE Index Average: 0.86, 1.00
  - Overall bilateral correlations lie in the 0.66-0.87 range; the Japanese yen is a notable exception with weaker correlations (driven by the pre-GFC sample).
- Aggregation and robustness:
  - Comparable results for shorter horizons (1-year and 3-months) are reported in Annex A (not reproduced here).
  - Annex A decomposition indicates most movements in λ_{AE,t} are associated with expected exchange rate changes rather than interest rate differentials.

### Interpretation and plausible drivers
- Possible risk-premium channel:
  - When risk appetite falls, the US dollar appreciates (a relatively safe asset), raising risk premia for other currencies; this generates (i) positive excess returns for foreign currency j over the US dollar and (ii) the positive correlation between UIP deviations and the USD index.
- Role of global financial intermediaries:
  - Higher demand for US dollars may lead intermediaries to demand a higher expected return for supplying dollars, contributing to UIP deviations.
- The ultimate sources of financial-market-driven US dollar fluctuations remain an active area of research and are beyond the present study.

### Comparison with other global financial indicators
- Correlations (quarterly levels, 1999:Q1–2022:Q4) with other measures:
  - USD Index with UIP Deviations: 0.86
  - USD Index with Global Financial Cycle: −0.39
  - UIP Deviations with Global Financial Cycle: −0.25
  - USD Index with VIX: 0.16
  - UIP Deviations with VIX: 0.25
  - USD Index with Commodity Price Index: −0.55
  - UIP Deviations with Commodity Price Index: −0.56
- Interpretation:
  - The correlation between the USD index and UIP deviations is considerably stronger and more robust than correlations between the USD index or UIP deviations and commonly studied measures of global financial conditions (Global Financial Cycle, VIX, Commodity Price Index).
  - The Global Financial Cycle correlation with the USD index weakens when the 2008-10 GFC period is excluded in levels, but quarterly-difference correlations remain well below the USD–UIP relation.
  - The VIX shows modest correlations (~0.2) with the USD index and UIP deviations over the sample; VIX’s correlation with the USD index is strong (0.73) only during the pre-GFC sub-period.
  - A global commodity price index exhibits negative correlation with both the USD index and UIP deviations (−0.56).

*wpiea2025065-print-pdf - 2.1    Constructing a Measure of UIP Deviations*

### 4.2    Policy Regimes and Structural Characteristics in Emerging Markets

### 4.2    Policy Regimes and Structural Characteristics in Emerging Markets

### Overview
- Purpose: Analyze how US dollar appreciations differentially affect economies based on policies and structural characteristics (monetary and exchange rate regimes, commodity dependence, trade openness, US dollar exposures).
- Method: State-dependent responses estimated via sample splits or country-specific factor values; factors captured as country-specific averages over the sample period to avoid endogeneity with the US dollar shock.
- Identification challenges:
  - Many characteristics correlate strongly with the EM vs AE sample split (notably monetary policy anchoring).
  - Country characteristics are often highly collinear with each other (e.g., exchange rate regimes correlate with US dollar invoicing; commodity exporters correlate with higher US dollar liabilities, lower GVC participation, higher USD invoicing, less flexible exchange rates, and lower trade openness).
- Strategy: Limit identification to heterogeneity within EMs, focusing first on commodity dependence (slow moving), then monetary policy anchoring, conditional exchange rate regime analysis, and other characteristics conditional on commodity dependence.

### Commodity Dependence
- Key finding:
  - Commodity exporters exhibit larger negative spillovers owing to concurrent deteriorations in their terms of trade.
- Exact quantified effects:
  - A 1 percent US dollar appreciation decreases the terms of trade by 0.7 percent after three quarters.
- Dynamics and mechanisms:
  - Commodity importers’ terms of trade remain broadly unaffected; their output decline is shallower and eventual REER depreciation buffers spillovers.
  - No evidence that the REER depreciates disproportionately for commodity exporters to offset falling commodity prices (consistent with fear of floating).
  - No evidence of accommodative monetary policy in commodity exporters; they tighten monetary policy significantly from the 3rd quarter onward.
- Contrast with advanced economies:
  - AE commodity exporters also show sizable negative terms-of-trade responses but do not experience more negative output impacts than AE commodity importers, partly because AE commodity exporters allow significant REER depreciation and pursue more accommodative monetary policy.

### Monetary Policy Credibility
- Key finding:
  - Monetary policy anchoring mitigates negative spillovers from US dollar appreciations by facilitating accommodative policy responses.
- Exact dynamics:
  - Emerging markets with more anchored inflation expectations exhibit a shallower initial decline in output (difference with less anchored EMs is statistically significant).
  - When inflation expectations are anchored:
    - REER depreciates on impact.
    - Policy rate becomes more accommodative (policy rates decrease).
    - Countries decrease current account balances to smooth temporary output drops.
  - In contrast, less anchored EMs:
    - Policy rates increase.
    - REER appreciates on impact, contributing to larger negative spillovers.
- Note:
  - Controlling for commodity dependence does not affect these findings, as terms of trade do not respond differently across anchoring groups.

### Exchange Rate Flexibility
- Key finding:
  - Freely floating exchange rate regimes facilitate faster recoveries in output for EMs after a US dollar appreciation.
- Exact dynamics:
  - Freely floating EMs: REER depreciates sizably on impact and remains depreciated into the medium term; avoid deflationary impact on CPI; credit recovers in medium term; private capital inflows outperform; current account decreases as output recovers.
  - Other (less flexible) regimes: REER does not depreciate during the first two years; persistent deflationary impulse on CPI; credit declines in medium term; current account tends to increase.
- Collinearity caveat:
  - Exchange rate regimes are highly correlated with the share of exports invoiced in US dollars (correlation of -0.9), linking freely floating regimes with lower USD export invoicing shares.
  - Findings for USD export invoicing mirror those for exchange rate flexibility:
    - Lower USD export invoicing share speeds up output recovery via REER depreciation.
    - High USD invoicing EMs tighten monetary policy, exhibit initial deflationary CPI impulse, and REER does not depreciate during the initial 8 quarters.
  - Collinearity prevents clean differentiation of the independent effects of exchange rate regime versus USD invoicing.

### Other Characteristics (conditional on commodity dependence)
- Trade openness:
  - EMs with high trade openness: negative output spillovers are shallower and limited to the initial three quarters (resembling AE impacts); REER depreciates on impact; sizable initial decrease in the current account balance smooths output impact.
  - EMs with low trade openness: protracted output fall extending to 10 quarters; significant monetary tightening and initial REER appreciation.
  - GVC participation is highly correlated with trade openness (+0.9) and shows similar impulse responses.
- US dollar liability exposure:
  - More USD-denominated foreign liabilities tend to be associated with harder hits from USD appreciations.
  - However, much of the significant negative output response is driven by commodity exporters that are also more USD-liability exposed.
  - Controlling for commodity dependence: EMs with higher USD liability exposure are only marginally more negatively affected in output than less exposed EMs.
  - Dynamics: Highly exposed EMs experience REER appreciation on impact and tighten monetary policy; low-exposure EMs depreciate REER in the short run and decrease policy rates.
- Reserve assets:
  - Countries with higher stocks of reserve assets are estimated to systematically intervene to buffer exchange rate movements.
  - No significant difference in short-run exchange rate or GDP responses to the shock based on reserve asset levels.

### Robustness
- Sensitivity checks maintained baseline results:
  - Adding various monetary policy shock series (one at a time) does not change baseline results.
  - Controlling for oil supply shocks from Baumeister and Hamilton (2019) yields similar results.
  - Dropping fixed exchange rate regimes from the sample does not drive the EM results.
  - Restricting the sample to the pre-COVID19 period (sample ends in 2019Q4) yields a stronger rebound in output one year after the shock but preserves output response differences between EMs and AEs.
  - Removing global controls increases the magnitude of output declines for both AEs and EMs, but AEs still experience a shallower, shorter-lived contraction.
  - Excluding all AEs included in the US dollar trade weighted index (Australia, Austria, Belgium, Finland, Greece, Netherlands, Portugal, Spain, and Sweden) does not materially change point estimates, though error bands increase.

### Conclusion and Policy Implications
- Aggregate finding:
  - Negative spillovers from US dollar appreciations are more pronounced in emerging market economies, with larger and longer-lived declines in output than in advanced economies.
  - REER depreciation facilitates adjustment in advanced economies; in EMs, REER does not adjust on impact and depreciates only gradually (consistent with fear of floating).
  - Financial channels (reduced capital inflows, decline in domestic credit) contribute to adverse spillovers in EMs.
- Role of commodity exposure:
  - Commodity export exposure magnifies spillovers because US dollar appreciations are historically associated with deteriorating commodity prices, producing terms-of-trade declines that are not offset by REER depreciation in EM commodity exporters.
- Policy recommendations for EMs:
  - Anchor inflation expectations to mitigate negative spillovers through accommodative monetary responses, REER depreciation, and policy rate decreases.
  - Adopt and maintain exchange rate flexibility to speed up economic recovery following US dollar appreciations.
- Broader research implication:
  - Understanding multilateral policies affecting the global dollar cycle requires deeper analysis of uncovered UIP deviations—linked to market-wide risk appetite and risk premia demanded by global financial intermediaries—which reflect intermediary frictions and spillovers from financial regulation.

*Source: IMF staff analysis, section 4.2 of the provided chapter.*

### References

### References

### Monetary policy, shocks, and identification
- Acosta, M. (2023). The Perceived Causes of Monetary Policy Surprises.Unpublished.
- Bauer, M. D. and E. T. Swanson (2023). A reassessment of monetary policy surprises and high-frequency identification.NBER Macroeconomics Annual 37, 87–155.
- Bu, C., J. Rogers, and W. Wu (2021). A unified measure of Fed monetary policy shocks.Journal of Monetary Economics 118(C), 331–349.
- Cloyne, J., Ò. Jordà, and A. M. Taylor (2023). State-dependent local projections: Understanding impulse response heterogeneity. Technical report, National Bureau of Economic Research.
- Dedola, L., G. Rivolta, and L. Stracca (2017). If the fed sneezes, who catches a cold?Journal of International Economics 108, S23–S41. 39th Annual NBER International Seminar on Macroeconomics.
- Di Giovanni, J. and J. Rogers (2024). The impact of us monetary policy on foreign firms.IMF Economic Review 72(1), 58–115.
- Dornbusch,  R. (1976).   Expectations and Exchange Rate Dynamics.Journal of Political Economy 84(6), 1161–1176.
- Eichenbaum, M. S., B. K. Johannsen, and S. T. Rebelo (2021).  Monetary Policy and the Predictability of Nominal Exchange Rates.The Review of Economic Studies 88(1), 192–228.
- Frenkel,  J. A. (1976).   Inflation and the formation of expectations.Journal of Monetary Economics 1(4), 403–421.

### Exchange rates, invoicing, and external balance
- Allen, C., D. Gautam, and L. Juvenal (2023). Currencies of External Balance Sheets. IMF Working Papers 2023/237, International Monetary Fund.
- Boz, E., C. Casas, G. Georgiadis, G. Gopinath, H. Le Mezo, A. Mehl, and T. Nguyen (2022).  Patterns of invoicing currency in global trade: New evidence.Journal of International Economics 136, 103604. NBER International Seminar on Macroeconomics 2021.
- Engel, C. and K. D. West (2005).  Exchange Rates and Fundamentals.Journal of Political Economy 113(3), 485–517.
- Engel, C. and S. P. Wu (2024). Exchange rate models are better than you think, and why they didn’t work in the old days. Working Paper 32808, National Bureau of Economic Research.
- Fukui, M., E. Nakamura, and J. Steinsson (2023).  The Macroeconomic Consequences of Exchange Rate Depreciations. NBER Working Papers 31279, National Bureau of Economic Research, Inc.
- Gabaix, X. and M. Maggiori (2015).  International Liquidity and Exchange Rate Dynamics.The Quarterly Journal of Economics 130(3), 1369–1420.
- Druck, P., N. E. Magud, and R. Mariscal (2018).  Collateral damage: Dollar strength and emerging mar-kets’ growth.The North American Journal of Economics and Finance 43(C), 97–117.
- Georgiadis, G., G. J. Müller, and B. Schumann (2023).  Global Risk and the Dollar.  Discussion Papers of DIW Berlin 2057, DIW Berlin, German Institute for Economic Research.
- Akinci, O., G. Benigno, S. Pelin, and J. Turek (2023).  The Dollar’s Imperial Circle.  Liberty Street Economics 20230301, Federal Reserve Bank of New York.
- Gournichas, P.-O. (2018).  Monetary Policy Transmission in Emerging Markets: An Application to Chile. In E. G. Mendoza, E. Pastén, and D. Saravia (Eds.),Monetary Policy and Global Spillovers:  Mechanisms, Effects and Policy Measures,  Volume 25 ofCentral Banking,  Analysis,  and Economic Policies Book Series,

### Global liquidity, cross-border banking, and spillovers
- Bruno, V. and H. S. Shin (2015).  Cross-Border Banking and Global Liquidity.Review of Economic Studies 82(2), 535–564.
- Di Giovanni, J.,  Kalemli-Ozcan, M. F. Ulu, and Y. S. Baskaya (2021).  International Spillovers and Local Credit Cycles.The Review of Economic Studies 89(2), 733–773.
- Bems, R., F. Caselli, F. Grigoli, and B. Gruss (2021).  Expectations’ anchoring and inflation persistence. Journal of International Economics 132(C).

### Models, methods, and structural analysis
- Andrle, M., P. Blagrave, P. Espaillat, M. K. Honjo, M. B. Hunt, M. Kortelainen, R. Lalonde, M. D. Laxton, E. Mavroeidi, M. D. Muir, et al. (2015).The flexible system of global models–FSGM. International Monetary Fund.
- Baumeister, C. and J. D. Hamilton (2019).  Structural Interpretation of Vector Autoregressions with In-complete Identification:  Revisiting the Role of Oil Supply and Demand Shocks.American Economic Review 109(5), 1873–1910.
- Cloyne, J., Ò. Jordà, and A. M. Taylor (2023). State-dependent local projections: Understanding impulse response heterogeneity. Technical report, National Bureau of Economic Research.
- De Marco, F., M. Macchiavelli, and R. Valchev (2021). Beyond home bias: International portfolio holdings and information heterogeneity.The Review of Financial Studies 35(9), 4387–4422.
- Bems, R., F. Caselli, F. Grigoli, and B. Gruss (2021).  Expectations’ anchoring and inflation persistence. Journal of International Economics 132(C).

### Data sets, indices, and applied measures
- Ahir, H., N. Bloom, and D. Furceri (2022). The world uncertainty index. Working Paper 29763, National Bureau of Economic Research.
- Adler, G., K. S. Chang, R. C. Mano, and Y. Shai (2024).   Foreign exchange intervention:  A data set of official data and estimates.Journal of Money, Credit and Banking.
- Bu, C., J. Rogers, and W. Wu (2021). A unified measure of Fed monetary policy shocks.Journal of Monetary Economics 118(C), 331–349.

*wpiea2025065-print-pdf - References*

### Chapter 8, pp. 279–324. Central Bank of Chile.

### Chapter 8. Central Bank of Chile

### A. Additional Results for UIP Deviations and the US Dollar Cycle
- Data and construction
  - UIP calculation implemented as in equation (1) on the specific date Consensus Economics exchange rate forecast survey is published each month by selecting interest rates and spot exchange rates for the corresponding day.
  - For Switzerland and Sweden, forecasts reported vs. the euro are converted to local currency per $ using the corresponding euro/$ forecast.
  - For euro/$ UIP deviations, relevant interest rates proxied with interest rates for Germany.
  - Interest rate types examined: (i) money market rates, (ii) deposit rates, (iii) yields on government’s local currency borrowing. Correlations between UIP deviations constructed from these three types exceed 0.99 at the 1-year horizon. Results reported use local currency government bond yields.

- Extended results and robustness
  - 1-year horizon:
    - UIP deviations centered around zero, mean deviation of 0.002.
    - Mean of absolute value of deviations is 0.020, which is 23% smaller than for the 2-year horizon (0.026).
    - UIP deviations are strongly correlated with the USD index; correlation is broad based across underlying AE currencies, though Japanese yen and British pound show weaker correlations.
  - 3-months horizon:
    - Deviations centered around zero at -0.005 and are 56% smaller than for the 2-year horizon.
    - UIP deviations at the 3-months horizon do not show a significant correlation with the USD index; aggregated measure correlation with USD index is 0.18.
    - Individual currency correlations at 3-months range from -0.38 to 0.57.

- Selected correlation statistics (preserve exactly as in source)
  - Table A1 (1-Year UIP deviations correlations):
    - GBP: LC/USD 0.27, USD index 0.46, λ_AE_t 0.75
    - SWK: LC/USD 0.59, USD index 0.66, λ_AE_t 0.87
    - CHF: LC/USD 0.82, USD index 0.56, λ_AE_t 0.85
    - CAD: LC/USD 0.67, USD index 0.67, λ_AE_t 0.74
    - JPY: LC/USD 0.40, USD index 0.10, λ_AE_t 0.31
    - EUR: LC/USD 0.71, USD index 0.65, λ_AE_t 0.95
    - AUD: LC/USD 0.70, USD index 0.62, λ_AE_t 0.87
    - AE Index Average – 0.69 1.00
  - Table A2 (3-Months UIP deviations correlations):
    - GBP: LC/USD -0.38, USD index 0.21, λ_AE_t 0.81
    - SWK: LC/USD 0.06, USD index 0.18, λ_AE_t 0.86
    - CHF: LC/USD 0.58, USD index 0.22, λ_AE_t 0.77
    - CAD: LC/USD 0.27, USD index 0.28, λ_AE_t 0.74
    - JPY: LC/USD -0.16, USD index -0.34, λ_AE_t 0.21
    - EUR: LC/USD 0.32, USD index 0.29, λ_AE_t 0.89
    - AUD: LC/USD 0.20, USD index 0.12, λ_AE_t 0.82
    - AE Index Average – 0.18 1.00
  - Table A3 (2-Year UIP deviations in first differences, ∆(λ_j/$_t) correlations):
    - GBP 0.66; SWK 0.76; CHF 0.80; CAD 0.85; JPY 0.68; EUR 0.77; AUD 0.84; AE Index Average – 0.80 1.00

- Decomposition of UIP deviations (equation A.1)
  - λ_AE,k_t = Σ_j ω_j [ln(1+i_j,k_t) − ln(1+i_$,k_t)] + Σ_j ω_j [ln(S_j/$_t) − ln(E(S_j/$_t+k))]
  - Findings: bulk of UIP deviations are related to the expected exchange rate adjustment; interest rate differentials have been relatively subdued and less volatile for advanced economies over the last 20 years.
  - Figure A2 confirms exchange rate adjustment dominates across major AE currencies.

- Correlations with other global financial condition measures (Table A4)
  - Levels (full sample):
    - USD Index vs. UIP Deviations 0.86
    - USD Index vs. Global Financial Cycle -0.39
    - USD Index vs. VIX 0.16
    - USD Index vs. World Uncertainty Index 0.14
    - USD Index vs. Commodity Price Index -0.55
  - First differences:
    - USD Index vs. UIP Deviations 0.80
    - USD Index vs. Global Financial Cycle -0.56
    - USD Index vs. VIX 0.35
    - USD Index vs. World Uncertainty Index 0.12
    - USD Index vs. Commodity Price Index -0.59
  - Pre-GFC and post-GFC breakdowns reported; level-wise correlations with USD Index are particularly strong for several measures in the pre-GFC period.

### B. Data
- Figure B3: Distribution of country characteristics shows share of AEs and EMs above/below full sample mean across characteristics: "Commodity exporter", "Less anchored inflation expectations", "High exposure to US dollar liabilities", "Low GVC participation", "High US dollar export invoicing", "Low trade openness".
  - Advanced economies exclude countries with weights in the US dollar index larger than 4 percent in 2020: Canada, France, Germany, Ireland, Italy, Japan, Switzerland, United Kingdom.
  - Countries that are not freely floating but are anchored to a currency other than the US dollar, that is freely floating against the US dollar, are classified as freely floating.

- Figure B4: Plots of average net commodity exports (share of GDP) vs. (i) average degree of inflation anchoring, (ii) average weight of the US dollar in foreign liabilities, (iii) average global value chain participation, (iv) average weight of US dollar invoicing for exports.

- Selected country-characteristic correlations and percentiles
  - Table B5: Pairwise correlations for EMs (select entries):
    - Commodity dependence with Monetary policy credibility 0.26
    - Commodity dependence with US dollar liability exposure 0.48
    - US dollar liability exposure with Exchange rate regime -0.57
    - US dollar export invoicing highly correlated with Exchange rate regime -0.95 and with USD liabilities 0.81
  - Table B6: Exchange rate regime and trade invoicing - lists of countries by regime and USD export invoicing share above/below average (as in source).
  - Table B7: Policies and structural features in EMs — percentiles and country averages (preserve exact numbers):
    - Net Commodity Exports: Min -0.06, Max 0.17, Mean 0.03, 10th Pct -0.05, 25th Pct -0.03, Median 0.01, 75th Pct 0.07, 90th Pct 0.17
    - MP Credibility: Min -1.38, Max 0.33, Mean -0.13, 10th Pct -0.95, 25th Pct -0.18, Median 0.03, 75th Pct 0.11, 90th Pct 0.20
    - Trade Openness: Min 26.39, Max 157.57, Mean 65.35, 10th Pct 28.21, 25th Pct 43.10, Median 52.41, 75th Pct 72.21, 90th Pct 150.80
    - USD Export Invoicing: Min 17.21, Max 98.76, Mean 70.85, 10th Pct 25.85, 25th Pct 47.02, Median 81.85, 75th Pct 94.43, 90th Pct 97.12
    - GVC Participation: Min 0.09, Max 0.45, Mean 0.20, 10th Pct 0.10, 25th Pct 0.12, Median 0.18, 75th Pct 0.26, 90th Pct 0.36
    - USD Liab. to Total Liab.: Min 0.09, Max 0.51, Mean 0.30, 10th Pct 0.12, 25th Pct 0.18, Median 0.33, 75th Pct 0.36, 90th Pct 0.45
    - FX Reserves: Min 0.05, Max 0.37, Mean 0.18, 10th Pct 0.09, 25th Pct 0.11, Median 0.16, 75th Pct 0.23, 90th Pct 0.36

- Data sources (Table B8) — selected indicators and sources as in text
  - Nominal US Dollar Trade-Weighted Index: Haver Analytics based on Fed’s Nominal Advanced Foreign Economies US Dollar Index (FRED).
  - Bilateral exchange rates: Haver Analytics.
  - Real effective exchange rates: IMF Information Notice System (INS).
  - Policy Rate: BIS Central Bank Policy Rates; Haver Analytics; IMF IFS.
  - UIP deviations: Consensus Economics; Refinitiv Datastream; Haver Analytics; Federal Reserve Board; IMF staff estimates.
  - Commodity trade balance: UN Comtrade; commodity list specified in note.
  - Share of external liabilities in US dollars: Allen et al. (2023).
  - Foreign exchange reserves: External Wealth of Nations Database Milesi-Ferretti (2024) based on Lane and Milesi-Ferretti (2018).
  - Notes on data frequency, seasonal adjustment, and specific imputations and assumptions provided in the source (see Table B8 notes).

### C. Additional Results (Econometric and Robustness)
- Instrument relevance and first-stage strength (Table C1)
  - First-stage regression of changes in the US dollar index on UIP deviations with controls from equation (3).
  - IV Coefficient range (Emerging Markets and Advanced Economies): coefficients significant at 1 percent.
  - Kleibergen-Paap Wald F-statistics reported: Min., Median, Max. values (examples):
    - Emerging Markets F-statistics: Min. 65.35, Median 159.66, Max. 254.70
    - Advanced Economies F-statistics: Min. 118.90, Median 164.68, Max. 194.81
  - Observations: 779, 595, 959, 595, 959, 595 (as reported in table formatting).

- Spillover heterogeneity (figures summarized)
  - Advanced economies: impulse responses to a 1% appreciation in the nominal US dollar AE Index, with responses for commodity importer vs exporter at 10th and 90th percentiles (net commodity imports 5% of GDP; net commodity exports 10% of GDP).
  - Emerging markets: impulse responses by exchange rate regime, by USD export invoicing (10th and 90th percentile as in Table B7), by GVC participation (10th and 90th percentile), by USD liabilities (10th and 90th percentile), and by reserve holdings (10th and 90th percentile).
  - Robustness checks:
    - Controlling for other shocks: shocks from Jarociński and Karadi (2020); Nakamura and Steinsson (2018) and Gürkaynak et al. (2005) as updated by Acosta (2023); Bu et al. (2021); Bauer and Swanson (2023); Baumeister and Hamilton (2019).
    - Excluding countries with fixed exchange rates and excluding COVID-19 period (sample ends in 2019q4) — results plotted with 90% and 68% confidence bands.

### D. Model Simulations: Flexible System of Global Models (FSGM)
- Model and calibration
  - Uses FSGM (Andrle et al., 2015), semistructural multiregion general equilibrium model; analysis uses G20MOD module covering every G20 economy.
  - Key features:
    - Monetary authorities follow an inflation-forecast-based rule under flexible exchange rate, with higher weight on exchange rate deviations for emerging markets.
    - Long-term (10-year) interest rate based on expectations theory plus a term premium; household/firm/government interest rates are weighted averages of 1- and 10-year rates.
    - UIP deviations modeled as risk premiums; sovereign risk premium affects all interest rates; corporate risk premium affects private sector rates.
    - Commodities modeled: oil, food, metals; priced in US dollars.
    - External sector: exports/imports determined by foreign and domestic activity and exchange rate; producer pricing assumed.

- Simulation setup and main results
  - Shock: global persistent 1 percentage point shock to the sovereign premium (applied excluding the United States).
  - Key model outcomes (Figure D1 and narration):
    - The sovereign premium shock generates a US dollar appreciation via increased demand for US dollars as risk-free returns on foreign bonds fall.
    - Short-term interest rates do not immediately change; risk premium increases; advanced economy central banks ease policy rates in response to higher financing costs, contributing to US dollar appreciation.
    - Financing costs increase, reducing domestic consumption and investment, leading to a fall in output outside the US. The fall is larger in emerging markets due to more limited exchange rate flexibility.
    - Commodity prices decline: a 1 percent appreciation in the US dollar is associated with a 2.3 percent decline in commodity prices at a one-year horizon in the simulation.
      - The US dollar pricing channel accounts for about 10 percent of the overall fall in the commodity price after one year in the model decomposition.
    - Global trade openness falls as investment drops and imports decline (high import propensity of investment goods).
    - Current account effects:
      - Commodity importers: terms-of-trade improvement raises real income; saving and investment responses broadly offset consumption effects but the current account increases, more so for EM commodity importers because of larger fall in investment.
      - Commodity exporters: opposing forces (higher cost of capital raising current account vs. falling export values lowering current account) broadly offset, leaving current account unchanged in the simulation.
  - Model caveats noted:
    - Omits detailed balance sheet mismatches and intermediary frictions; does not model some financial spillovers explicitly (some captured via exogenous shock to financial conditions).
    - FSGM lacks financial intermediaries; UIP deviations serve as a proxy for risk premiums rather than capturing intermediary rents.

*Source: Chapter 8, Central Bank of Chile, pp. 279–324.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025065-print-pdf.pdf_