## ch2onlineannexes

## Source details

**Canonical URL:** [ch2onlineannexes](https://www.imf.org/-/media/files/publications/weo/2025/october/english/ch2onlineannexes.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/weo/2025/october/english/ch2onlineannexes.pdf.md)
- [Structured JSON version](/-/media/files/publications/weo/2025/october/english/ch2onlineannexes.pdf.json)

---

### Monetary policy: actual vs perceived reaction functions (EMs, pre‑GFC)
- Actual monetary policy reaction function (pre‑GFC = 1997–2009):
  - Inflation Gap: 2.251*** (0.806)
  - Output Gap: -0.104 (0.626)
  - NEER Appreciation: -0.514*** (0.121)
  - Observations: 3,455
  - R2: 0.993
- Perceived monetary policy reaction function (yearly estimates of one‑year ahead three‑month saving rate forecasts regressed on next‑year inflation expectations and real GDP growth forecasts; years with ≥1,000 observations):
  - 2004: Expected Inflation 2.360 (0.440); Expected Real GDP Growth -0.120 (0.160); R2 0.900; Observations 1,009
  - 2005: 2.800 (0.810); 0.260 (0.210); R2 0.870; Observations 1,189
  - 2006: 1.550 (0.450); 0.460 (0.210); R2 0.880; Observations 1,311
  - 2007: 2.610 (0.190); 0.220 (0.180); R2 0.880; Observations 1,604
  - 2008: 2.990 (0.150); -0.350 (0.120); R2 0.900; Observations 1,951
  - 2009: 2.080 (0.220); -0.180 (0.120); R2 0.890; Observations 2,123
  - 2010: 2.270 (0.230); -0.210 (0.160); R2 0.920; Observations 2,163
  - 2011: 2.390 (0.190); -0.450 (0.150); R2 0.910; Observations 2,251
  - 2012: 2.470 (0.170); -0.950 (0.200); R2 0.950; Observations 2,185
  - 2013: 2.110 (0.230); -0.710 (0.150); R2 0.940; Observations 2,139
  - 2014: 2.410 (0.210); -0.590 (0.100); R2 0.940; Observations 2,509
  - 2015: 1.670 (0.120); -0.240 (0.100); R2 0.920; Observations 2,695
  - 2016: 1.100 (0.060); -0.070 (0.070); R2 0.820; Observations 2,736
  - 2017: 0.970 (0.130); -0.020 (0.220); R2 0.900; Observations 2,882
  - 2018: 1.310 (0.180); 0.440 (0.240); R2 0.910; Observations 2,818
  - 2019: 1.350 (0.140); 0.540 (0.150); R2 0.870; Observations 2,976
  - 2020: 1.900 (0.150); 0.160 (0.270); R2 0.790; Observations 2,733
  - 2021: 1.880 (0.120); 0.210 (0.180); R2 0.820; Observations 2,703
  - 2022: 1.510 (0.140); 0.020 (0.060); R2 0.880; Observations 2,548
  - 2023: 1.240 (0.080); 0.310 (0.110); R2 0.790; Observations 2,586
  - 2024: 1.330 (0.120); 0.580 (0.140); R2 0.830; Observations 2,713
- Key finding:
  - Perceived reaction‑function coefficients show a decline in the perceived weight on expected inflation over time.

### Fiscal dominance: pre‑GFC vs post‑GFC
- Estimation: regressions of two‑year ahead inflation expectations and next‑year policy rate on previous year military spending growth, with controls and country & year fixed effects.
- Responses to fiscal spending shocks:
  - Fiscal Spending Shock (EMs): Expected Inflation 0.024** (0.010); Policy Rates -0.099** (0.041)
  - Fiscal Spending Shock (AEs): Expected Inflation 0.001 (0.001); Policy Rates 0.003 (0.004)
  - Fiscal Spending Shock × Post‑GFC (EMs): Expected Inflation -0.028* (0.014); Policy Rates 0.124** (0.050)
  - Country Fixed Effects: Yes; Time Fixed Effects: Yes
  - Observations: EMs 468/438; AEs 610/615
  - Adjusted R2: EMs 0.849/0.833; AEs 0.800/0.924
- Findings:
  - Pre‑GFC: spending increases followed by monetary easing and higher inflation expectations — evidence consistent with fiscal dominance.
  - Post‑GFC: central banks increased policy rates after spending increases (coefficients borderline insignificant in SVAR), and long‑run inflation expectations remained close to target — dynamics resemble advanced economies.

### SVAR analysis of primary spending shocks
- System specification and identification:
  - Variables: log GDP, exchange rate depreciation, policy rates, 5‑year ahead inflation expectations, log government primary expenditure
  - Lags k = 3; identification via Cholesky ordering: log GDP → exchange depreciation → policy rates → 5‑year ahead inflation expectations → log government primary expenditure
- Impulse responses:
  - Pre‑GFC: a one percent increase in primary expenditure led to monetary easing and increased expected inflation.
  - Post‑GFC: a one percent increase in primary expenditure led to central banks increasing policy rates (estimates borderline insignificant), leaving long‑run inflation expectations unchanged.

### Monetary policy autonomy and transmission to financial markets
- One‑day asset responses to monetary policy shocks (coefficients with robust standard errors):
  - 3 month yield:
    - Domestic Monetary Policy Shock: 0.104*** (0.031)
    - US Monetary Policy Shock: 0.024 (0.017)
  - 10 year yield:
    - Domestic Monetary Policy Shock: 0.028*** (0.008)
    - US Monetary Policy Shock: 0.021*** (0.006)
  - Stock Price:
    - Domestic Monetary Policy Shock: -0.090** (0.037)
    - US Monetary Policy Shock: -0.244*** (0.047)
  - Nominal Exchange Rate:
    - Domestic Monetary Policy Shock: -0.066* (0.038)
    - US Monetary Policy Shock: 0.148*** (0.030)
  - EMBI Spread:
    - Domestic Monetary Policy Shock: 0.122 (0.116)
    - US Monetary Policy Shock: 0.569*** (0.139)
  - Country Fixed Effects: Yes
  - Observations by instrument: 2,610; 4,063; 5,845; 6,131; 5,261
  - Adjusted R2: 0.014; 0.014; 0.008; 0.010; 0.003
- Interpretation:
  - Domestic monetary policy shocks transmit strongly to government bond yields, especially at the short end.
  - US monetary policy shocks have smaller pass‑through to yields but larger effects on stock prices, exchange rates, and credit spreads; effects on 10‑year yields are broadly comparable to domestic shocks.

### FX interventions and policy frameworks
- Local projection of cumulative FX interventions (net sales relative to GDP) in response to 12‑month UIP deviations, instrumented with the RORO Index; interaction with policy‑framework proxies.
- Key interaction coefficients (12 months after a one percentage point UIP deviation):
  - Inflation Expectations Anchoring × UIP: -1.214*** (0.376)
  - Macroprudential Net Tightening × UIP: -0.353*** (0.127)
  - Country Fixed Effects: Yes; Time Fixed Effects: Yes
  - Observations: 2,032 / 2,379
  - Adjusted R2: -0.143 / -0.128
- Findings:
  - EMs with well‑anchored inflation expectations intervene less (smaller net FX sales) in response to UIP deviations triggered by risk‑off episodes.
  - EMs with stronger macroprudential frameworks (net cumulative tightening in FX‑related capital requirements, loan and position restrictions) are more likely to allow UIP deviations to play out rather than counteract them via FX sales.

### Fiscal policy frameworks: credibility, cyclicality, and reaction functions
- Anchoring of fiscal expectations (private forecasts regressed on official forecasts; April submissions):
  - Current Year Forecast, Official:
    - EMs Pre‑GFC: 0.669*** (0.072)
    - EMs Post‑GFC: 0.825*** (0.076)
    - AEs: 0.788*** (0.036)
  - Planned Adjustment, Official:
    - EMs Pre‑GFC: 0.389*** (0.032)
    - EMs Post‑GFC: 0.694*** (0.063)
    - AEs: 0.566*** (0.056)
  - Country Fixed Effects: Yes
  - Observations: EMs 122 / 211 / 431; AEs 99 / 174 / 335
  - R2: 0.896; 0.782; 0.918; 0.665; 0.446; 0.586
- Interpretation:
  - Private sector forecasts of fiscal balance and planned adjustments have become more closely anchored to official forecasts post‑GFC, indicating increased fiscal credibility.

- Cyclicality of fiscal policy (change in primary government expenditures/GDP on change in output gap with POST dummy and interactions; controls include lagged public debt/GDP):
  - Change in Output Gap (All Crisis Years): -0.087 (0.054)
  - Change in Output Gap × Post‑GFC (All Crisis Years): -0.242*** (0.074)
  - Change in Output Gap (Non‑Crisis Years): -0.106 (0.067)
  - Change in Output Gap × Post‑GFC (Non‑Crisis Years): -0.357*** (0.094)
  - Change in Output Gap (Commodity Importers): 0.136* (0.081)
  - Change in Output Gap × Post‑GFC (Commodity Importers): -0.278** (0.111)
  - Change in Output Gap (Commodity Exporters): -0.06 (0.059)
  - Change in Output Gap × Post‑GFC (Commodity Exporters): -0.312*** (0.082)
  - Country Fixed Effects: Yes
  - Observations: 625 / 312 / 313 / 370 / 255 / 825
  - R2: 0.143 / 0.235 / 0.141 / 0.157 / 0.147 / 0.352
- Findings:
  - Countercyclicality of fiscal policy strengthened post‑GFC, particularly after risk‑off events and in both commodity importers and exporters.

- Fiscal reaction functions (Bohn‑style; Online Annex Table 2.5.3 and 2.5.3 notes):
  - Baseline regressions: primary balance/GDP on lagged public debt/GDP or lagged interest bill/GDP, with output gap, unemployment controls, country & time fixed effects.
  - AEs reported coefficients (selected):
    - Lagged Public Debt: 0.043*** (0.010) in column (1); 0.031** (0.013) in column (2); 0.036*** (0.006) in column (3)
    - Lagged Public Debt × Post‑GFC: 0.028** (0.011) in column (4)
    - Lagged Public Debt × Fiscal Rule: 0.034*** (0.013) in column (5)
    - Interest Bill: 0.499*** (0.091) in column (1); 0.417*** (0.159) in column (2); 0.774*** (0.118) in column (3)
    - Interest Bill × Post‑GFC: 0.478*** (0.124) in column (4)
    - Interest Bill × Fiscal Rule: 0.487*** (0.173) in column (5)
    - Country Fixed Effects: Yes; Time Fixed Effects: Yes
    - Observations by column: 588, 422, 835, 612, 438, 835
    - R2 by column: 0.566, 0.533, 0.629, 0.576, 0.534, 0.633
  - Note: EMs panel indicated separately; significance notation *** p<0.01; ** p<0.05; * p<0.1.
  - Key interpretation: emerging markets with fiscal rules have more aggressive reaction functions.

### Sensitivity of sovereign spreads to debt burdens
- Regressions of log EMBI spreads on public debt/GDP and external debt/exports (Online Annex Table 2.5.4):
  - Public Debt: 0.019*** (0.002) in column (1)
  - Public Debt × Post‑GFC: -0.005* (0.003) in column (1)
  - External Debt: 0.004*** (0.000) in column (2)
  - External Debt × Post‑GFC: -0.001** (0.001) in column (2)
  - Country Fixed Effects: Yes
  - Observations: 516 (public debt), 528 (external debt)
  - R2: 0.671 (public debt), 0.646 (external debt)

### Fiscal consolidation speed — half‑life simulations
- Simulation objective: time implied by estimated fiscal reaction functions to reduce half of an unexpected increase in public debt/GDP.
- Scenarios:
  - Two interest‑growth differentials: low (r−g = zero percent) and high (r−g = 2 percent)
- Debt path equation using estimated fiscal response β̂:
  - d_t − d̅ = (d_{t−1} − d̅) (1 + (r−g) − β̂)
- Initial conditions: d_0 = d̅; d_1 = d_0 + ε (ε positive shock)
- Implied half‑lives do not depend on levels of d̅ or ε.
- Reference: chapter Figure 2.9.3 plots illustrative simulations.

### Drivers of emerging market resilience (Online Annex 2.6)
- Regression for predicted cumulative change in log real GDP growth and log CPI 12 months after the start of risk‑off episode includes seven pre‑determined policy variables and pre‑trend controls.
- Policy determinants (selected coefficients; policy variables entered one by one, expressed in standard deviations):
  - Anchoring of Inflation Expectations:
    - Real GDP Growth: 0.001 (0.005)
    - Consumer Price Inflation: -0.023** (0.009)
  - Reserve Adequacy:
    - Real GDP Growth: 0.006** (0.003)
    - Consumer Price Inflation: -0.005 (0.004)
  - FX Mismatches:
    - Real GDP Growth: -0.009*** (0.003)
    - Consumer Price Inflation: 0.009** (0.005)
  - Macroprudential Policy:
    - Real GDP Growth: 0.005* (0.003)
    - Consumer Price Inflation: -0.005** (0.003)
  - External Debt Burden:
    - Real GDP Growth: -0.008*** (0.002)
    - Consumer Price Inflation: 0.011*** (0.004)
  - Cyclically‑adjusted Balance:
    - Real GDP Growth: 0.004 (0.002)
    - Consumer Price Inflation: -0.006* (0.003)
  - Episode Fixed Effects: Yes
  - Observations by column: 273, 347, 270, 378, 378, 344, 273, 347, 270, 378, 378, 344
  - R2 (selected): 0.377, 0.403, 0.404, 0.362, 0.380, 0.360, 0.632, 0.64, 0.658, 0.56, 0.568, 0.603
- External conditions regressions (Online Annex Table 2.6.2; common component of 12‑month changes):
  - Real GDP Growth, AEs: 0.010 (0.006) for Real GDP Growth; 0.020** (0.007) for Consumer Price Inflation
  - Commodity Terms of Trade: 0.016* (0.009) for Real GDP Growth; 0.021* (0.010) for Consumer Price Inflation
  - Financial Conditions Impulse on Growth (FCI‑G): -0.002 (0.009) for Real GDP Growth; 0.033*** (0.010) for Consumer Price Inflation
  - Observations: 16 (both columns)
  - R2: 0.488 (Real GDP Growth), 0.539 (Consumer Price Inflation)
- Contributions computation:
  - Policy frameworks contribution: Σ β̂_k ΔP̅_{i,k}^{τ} over significant policy variables, where ΔP̅_{i,τ−1}^k is the mean of the median value of policy variable k during post‑GFC episodes relative to pre‑GFC.
  - External conditions contribution: Σ β̂_m G_τ^m using external proxies (advanced economy real GDP growth relative to trend; commodity terms‑of‑trade shocks; US financial conditions), all expressed as cumulative changes during 12 months following episode start.
- Robustness:
  - Excluding individual episodes one at a time (including taper tantrum starting June 2013) and imputing missing policy variables using median values during a specific episode; Online Annex Figure 2.6.1 reports distributions of estimated contributions for the median EM.

### The Q‑IPF model: structure, calibration, and simulations (Online Annex 2.7)
- Purpose: Quantitative Integrated Policy Framework extends a two‑country New Keynesian model for small open economies to jointly consider monetary, FX, and macroprudential policies with financial frictions, exchange rate pass‑through, and occasionally binding external debt limits.
- Financial market structure highlights:
  - Financiers intermediate cross‑border borrowing, trade FX, absorb currency risk, hold symmetric nominal positions in domestic and foreign bonds; finite risk‑bearing capacity generates UIP risk premium governed by FX market depth.
  - Banks transfer domestic currency between financiers and households, face borrowing constraint proportional to nominal economy size, generating a positive credit spread when constrained.
  - Portfolio investors take positions in home‑currency bonds financed by issuing foreign‑currency bonds.
- Households and firms:
  - Households: trade local currency bonds paying nominal interest I_b, supply differentiated labor with Dixit‑Stiglitz aggregation, set wages in staggered Calvo style, and face endogenous inflation expectations anchoring (degree of anchoring inversely related to inflation deviation from target).
  - Firms: monopolistic competition, Cobb‑Douglas production, staggered Calvo price setting, with some firms indexing to past prices and inflation.
- Policy behavior:
  - Monetary policy: standard Taylor rule for short‑term rate I; central bank minimizes inflation and output deviations; can conduct sterilized FX interventions (non‑negative foreign bond positions) with fiscal authority bearing financial outcomes.
  - Fiscal policy: public consumption exogenous; government balances budget with lump sum taxes.
- FX mismatches:
  - Determined by fraction of financiers and portfolio investors owned by domestic households and share of external liabilities denominated in domestic currency.
- Calibration:
  - Calibrated to two representative small open EMs: pre‑GFC (1997–2009) and post‑GFC (2010–24).
  - Post‑GFC calibration differences: better anchored inflation expectations; smaller balance sheet mismatches (higher net foreign assets and larger share of external liabilities in domestic currency); other parameters based on Adrian and others (2021).
- Exchange rate pass‑through estimation (Online Annex Table 2.7.1):
  - Cumulative CPI response to a one percentage point nominal exchange rate depreciation vis‑à‑vis the USD:
    - t + 6: FX Depreciation 0.332*** (0.121); FX Depreciation × Post‑GFC -0.319** (0.144)
    - t + 12: FX Depreciation 0.704*** (0.206); FX Depreciation × Post‑GFC -0.529** (0.258)
    - t + 24: FX Depreciation 1.358*** (0.403); FX Depreciation × Post‑GFC -1.189** (0.510)
    - t + 36: FX Depreciation 1.767*** (0.608); FX Depreciation × Post‑GFC -1.956** (0.787)
  - Country Fixed Effects: Yes; Time Fixed Effects: Yes
  - Observations: 5,291; 5,171; 4,931; 4,691
  - Adjusted R2: 0.662; 0.672; 0.682; 0.720
  - Significance: *** p<0.01; ** p<0.05; * p<0.1
- Model simulations and policy insights:
  - Simulations illustrate policy tradeoffs and mix/timing of responses across EMs with varying policy frameworks.
  - Frequency of sudden stops estimated with reduced financial frictions via stochastic simulations.
  - Pre‑GFC external borrowing constraint set so economy is constrained ≈ 3 percent of the time.
  - Simulations keep distance between steady‑state external debt and its limit constant across calibrations.

### Balance‑sheet and macro volatility statistics (Online Annex Table 2.8.1)
- Key statistics (Pre‑GFC = 1997–2009; Post‑GFC = 2010–24; AEs = advanced economies):
  - SD(ln Yt): Pre‑GFC 2.75; Post‑GFC 2.01; AEs 1.69
  - SD(Δln Yt): Pre‑GFC 1.87; Post‑GFC 1.37; AEs 1.04
  - SD(ln Ct): Pre‑GFC 2.71; Post‑GFC 2.53; AEs 1.52
  - SD(TBt): Pre‑GFC 4.33; Post‑GFC 4.27; AEs 3.54
  - ρ(ln Yt, ln Yt‑1): Pre‑GFC 0.74; Post‑GFC 0.76; AEs 0.80
  - ρ(Δln Yt, Δln Yt‑1): Pre‑GFC 0.08; Post‑GFC -0.08; AEs 0.06
  - SD(ln Ct) / SD(ln Yt): Pre‑GFC 1.01; Post‑GFC 1.25; AEs 0.89
  - ρ(TBt, ln Yt): Pre‑GFC -0.29; Post‑GFC -0.17; AEs -0.06
- Summary interpretation:
  - Pre‑GFC EMs exhibited higher income volatility, more volatile consumption relative to output, and a more counter‑cyclical trade balance than AEs.
  - Post‑GFC: income shocks in EMs became less pronounced and trade balance less procyclical, though differences with AEs remain.

*Source: ch2onlineannexes (Online annexes and tables accompanying IMF World Economic Outlook chapter materials).*

### 1. EMs,Pre-GFC

### 1. EMs, Pre-GFC

### Monetary policy: Actual and perceived reaction functions
- Actual monetary policy reaction function (Online Annex Table 2.3.1; pre-GFC = 1997–2009, post-GFC = 2010–24). Reported coefficients (EMs):
  - Inflation Gap: 2.251*** (0.806) [Pre-GFC]
  - Output Gap: -0.104 (0.626) [Pre-GFC]
  - NEER Appreciation: -0.514*** (0.121) [Pre-GFC]
  - Observations: 3,455
  - R2: 0.993
- Actual reaction function shows a progressive decline over time in the weight on expected inflation and a marginal increase in the weight on the output gap (interpreted as gains in monetary policy credibility).
- Perceived monetary policy reaction function (Online Annex Table 2.3.2; regressions of one-year ahead three‑month saving rate forecasts on next‑year inflation expectations and real GDP growth forecasts; years with ≥1,000 observations). Selected yearly estimates (EMs):
  - 2004: Expected Inflation, Coefficient 2.360; Expected Inflation, Standard Error 0.440; Expected Real GDP Growth, Coefficient -0.120; Expected Real GDP Growth, Standard Error 0.160; R2 0.900; Observations 1,009
  - 2005: 2.800; 0.810; 0.260; 0.210; 0.870; 1,189
  - 2006: 1.550; 0.450; 0.460; 0.210; 0.880; 1,311
  - 2007: 2.610; 0.190; 0.220; 0.180; 0.880; 1,604
  - 2008: 2.990; 0.150; -0.350; 0.120; 0.900; 1,951
  - 2009: 2.080; 0.220; -0.180; 0.120; 0.890; 2,123
  - 2010: 2.270; 0.230; -0.210; 0.160; 0.920; 2,163
  - 2011: 2.390; 0.190; -0.450; 0.150; 0.910; 2,251
  - 2012: 2.470; 0.170; -0.950; 0.200; 0.950; 2,185
  - 2013: 2.110; 0.230; -0.710; 0.150; 0.940; 2,139
  - 2014: 2.410; 0.210; -0.590; 0.100; 0.940; 2,509
  - 2015: 1.670; 0.120; -0.240; 0.100; 0.920; 2,695
  - 2016: 1.100; 0.060; -0.070; 0.070; 0.820; 2,736
  - 2017: 0.970; 0.130; -0.020; 0.220; 0.900; 2,882
  - 2018: 1.310; 0.180; 0.440; 0.240; 0.910; 2,818
  - 2019: 1.350; 0.140; 0.540; 0.150; 0.870; 2,976
  - 2020: 1.900; 0.150; 0.160; 0.270; 0.790; 2,733
  - 2021: 1.880; 0.120; 0.210; 0.180; 0.820; 2,703
  - 2022: 1.510; 0.140; 0.020; 0.060; 0.880; 2,548
  - 2023: 1.240; 0.080; 0.310; 0.110; 0.790; 2,586
  - 2024: 1.330; 0.120; 0.580; 0.140; 0.830; 2,713
- Perceived reaction-function coefficients reveal a decline in the perceived weight on expected inflation over time.

### Fiscal dominance (pre-GFC vs post-GFC)
- Estimation framework: regressions of two‑year ahead inflation expectations and next‑year policy rate on previous year military spending growth, with controls and country & year fixed effects.
- Responses of inflation expectations and policy rates to fiscal spending shocks (Online Annex Table 2.3.3):
  - Fiscal Spending Shock (EMs): Expected Inflation 0.024** (0.010); Policy Rates -0.099** (0.041)
  - Fiscal Spending Shock (AEs): Expected Inflation 0.001 (0.001); Policy Rates 0.003 (0.004)
  - Fiscal Spending Shock × Post‑GFC (EMs): Expected Inflation -0.028* (0.014); Policy Rates 0.124** (0.050)
  - Country Fixed Effects: Yes; Time Fixed Effects: Yes
  - Observations: EMs 468/438; AEs 610/615
  - Adjusted R2: EMs 0.849/0.833; AEs 0.800/0.924
- Findings:
  - Pre-GFC: increases in spending were followed by monetary easing and higher inflation expectations—evidence consistent with fiscal dominance.
  - Post-GFC: central banks increased policy rates after spending increases (coefficients borderline insignificant in SVAR), and long‑run inflation expectations remained close to target—post‑GFC dynamics resemble advanced economies.

### SVAR analysis of primary spending shocks
- System: variables = log GDP, exchange rate depreciation, policy rates, 5‑year ahead inflation expectations, log government primary expenditure; k=3 lags; identification via Cholesky decomposition ordering log GDP → exchange depreciation → policy rates → 5‑year ahead inflation expectations → log government primary expenditure.
- Impulse responses (Online Annex Figure 2.3.1 described):
  - Pre‑GFC: a one percent increase in primary expenditure led to monetary easing and increased expected inflation.
  - Post‑GFC: after a one percent increase in primary expenditure central banks increased policy rates (estimates borderline insignificant), leaving long‑run inflation expectations unchanged.

### Monetary policy autonomy and transmission to financial markets

### Responses to domestic and US monetary policy shocks (Online Annex Table 2.3.4)
- Estimated equation: one‑day changes in government bond yields, stock prices, spreads, and exchange rates regressed on domestic monetary policy shock and US monetary policy shock; country fixed effects included.
- Government bond yields / asset responses (coefficients and robust standard errors):
  - 3 month yield:
    - Domestic Monetary Policy Shock: 0.104*** (0.031)
    - US Monetary Policy Shock: 0.024 (0.017)
  - 10 year yield:
    - Domestic Monetary Policy Shock: 0.028*** (0.008)
    - US Monetary Policy Shock: 0.021*** (0.006)
  - Stock Price:
    - Domestic Monetary Policy Shock: -0.090** (0.037)
    - US Monetary Policy Shock: -0.244*** (0.047)
  - Nominal Exchange Rate:
    - Domestic Monetary Policy Shock: -0.066* (0.038)
    - US Monetary Policy Shock: 0.148*** (0.030)
  - EMBI Spread:
    - Domestic Monetary Policy Shock: 0.122 (0.116)
    - US Monetary Policy Shock: 0.569*** (0.139)
  - Country Fixed Effects: Yes across specifications
  - Observations vary by instrument: 2,610; 4,063; 5,845; 6,131; 5,261
  - Adjusted R2: 0.014; 0.014; 0.008; 0.010; 0.003
- Interpretation:
  - Domestic monetary policy shocks transmit strongly to government bond yields, especially at the short end of the yield curve, indicating retained monetary traction on borrowing conditions.
  - US monetary policy shocks have a considerably smaller pass‑through to yields but larger effects on riskier asset classes (stock prices, exchange rates, credit spreads); effects on 10‑year yields are broadly comparable to domestic shocks.

### FX interventions and policy frameworks

### Use of FX interventions in response to UIP deviations (Online Annex 2.4 and Table 2.4.1)
- Local projection specification for cumulative FX interventions (net sales relative to GDP) in response to 12‑month UIP deviations instrumented with the RORO Index; interaction with policy‑framework proxies.
- Key interaction coefficients (12 months after a one percentage point increase in the UIP deviation):
  - Inflation Expectations Anchoring × UIP: -1.214*** (0.376)
  - Macroprudential Net Tightening × UIP: -0.353*** (0.127)
  - Country Fixed Effects: Yes; Time Fixed Effects: Yes
  - Observations: 2,032 / 2,379
  - Adjusted R2: -0.143 / -0.128
- Findings:
  - Emerging markets with well‑anchored inflation expectations intervene less (net sales of FX are smaller) in response to UIP deviations triggered by risk‑off episodes.
  - Emerging markets with stronger macroprudential policy frameworks (net cumulative tightening in FX‑related capital requirements, loan and position restrictions) are more likely to allow UIP deviations to play out rather than counteracting them via foreign currency sales.

### Fiscal policy frameworks: credibility, cyclicality, and reaction functions

### Anchoring of fiscal expectations (Online Annex Table 2.5.1)
- Regression of private sector forecasts on official forecasts (April submissions).
- Selected coefficients:
  - Current Year Forecast, Official (EMs Pre‑GFC / Post‑GFC / AEs): 0.669*** (0.072); 0.825*** (0.076); 0.788*** (0.036)
  - Planned Adjustment, Official (EMs Pre‑GFC / Post‑GFC / AEs): 0.389*** (0.032); 0.694*** (0.063); 0.566*** (0.056)
  - Country Fixed Effects: Yes
  - Observations: EMs 122 / 211 / 431; AEs 99 / 174 / 335
  - R2: 0.896; 0.782; 0.918; 0.665; 0.446; 0.586
- Interpretation:
  - Private sector forecasts of the fiscal balance and planned adjustments have become more closely anchored to official forecasts post‑GFC, indicating increased fiscal credibility.

### Cyclicality of fiscal policy (Online Annex Table 2.5.2)
- Regression: change in primary government expenditures to GDP on change in the output gap, with POST dummy (2010 onwards), and interaction term; controls include lagged public debt to GDP and country fixed effects.
- Selected coefficients:
  - Change in Output Gap (All Crisis Years): -0.087 (0.054)
  - Change in Output Gap × Post‑GFC (All Crisis Years): -0.242*** (0.074)
  - Change in Output Gap (Non‑Crisis Years): -0.106 (0.067)
  - Change in Output Gap × Post‑GFC (Non‑Crisis Years): -0.357*** (0.094)
  - Change in Output Gap (Commodity Importers): 0.136* (0.081)
  - Change in Output Gap × Post‑GFC (Commodity Importers): -0.278** (0.111)
  - Change in Output Gap (Commodity Exporters): -0.06 (0.059)
  - Change in Output Gap × Post‑GFC (Commodity Exporters): -0.312*** (0.082)
  - Country Fixed Effects: Yes
  - Observations: 625 / 312 / 313 / 370 / 255 / 825
  - R2: 0.143 / 0.235 / 0.141 / 0.157 / 0.147 / 0.352
- Findings:
  - Improvements in countercyclicality are most pronounced in the years following risk‑off events and are present in both commodity importers and commodity exporters.
  - The negative and significant Change in Output Gap × Post‑GFC coefficients indicate stronger countercyclical fiscal responses post‑GFC.

### Fiscal reaction functions and responsiveness to debt sustainability pressures
- Baseline fiscal reaction function estimated à la Bohn (1998):
  - Primary balance to GDP (pb_i,t) regressed on lagged public debt to GDP or lagged interest bill to GDP (z_i,t−1), POST dummy (2010 onward), interaction POST × z_i,t−1, output gap and unemployment controls, and fixed effects.
- Alternate specification includes fiscal rule dummy (FR_i,t) and interaction FR_i,t × z_i,t−1 to assess differences between countries with and without fiscal rules.
- Overall interpretation:
  - The analysis assesses whether fiscal policy has become more responsive to debt sustainability pressures post‑GFC and whether fiscal rules strengthen responsiveness.

*Source: ch2onlineannexes - 1. EMs,Pre-GFC; International Monetary Fund | October 2025.*

### Annex Table 2.5.3 reports the results of the estimations. Columns (1), (3), (4) and (6) correspond

### Online Annex Tables and Model Details (ch2onlineannexes)

### Fiscal Reaction Functions (Online Annex Table 2.5.3)
- Regression specification: primary balance to GDP ratio on lagged interest bill (interest expenses to GDP ratio) and lagged public debt to GDP ratio, controlling for output gap, unemployment rate, country and time fixed effects. Robust standard errors in parentheses.
- AEs (advanced economies) — reported coefficients:
  - Lagged Public Debt: 0.043*** (0.010) in column (1)
  - Lagged Public Debt: 0.031** (0.013) in column (2)
  - Lagged Public Debt: 0.036*** (0.006) in column (3)
  - Lagged Public Debt × Post-GFC: 0.028** (0.011) in column (4)
  - Lagged Public Debt × Fiscal Rule: 0.034*** (0.013) in column (5)
  - Interest Bill: 0.499*** (0.091) in column (1)
  - Interest Bill: 0.417*** (0.159) in column (2)
  - Interest Bill: 0.774*** (0.118) in column (3)
  - Interest Bill × Post-GFC: 0.478*** (0.124) in column (4)
  - Interest Bill × Fiscal Rule: 0.487*** (0.173) in column (5)
  - Country Fixed Effects: Yes (all columns)
  - Time Fixed Effects: Yes (all columns)
  - Observations by column: 588, 422, 835, 612, 438, 835
  - R2 by column: 0.566, 0.533, 0.629, 0.576, 0.534, 0.633
- EMs (emerging markets) — table labels indicate separate panel for EMs; table notes:
  - AEs = advanced economies; EMs = emerging markets; GFC = global financial crisis.
  - Significance: *** p<0.01; ** p<0.05; * p<0.1.
- Key interpretation from columns (2) and (5): emerging markets with fiscal rules have more aggressive reaction functions (as indicated in the main text description).

### Sensitivity of Sovereign Spreads to Debt Burdens (Online Annex Table 2.5.4)
- Regression: log of sovereign EMBI spreads on public debt to GDP ratio or external debt to exports ratio, controlling for output gap, unemployment rate, and country fixed effects. Robust standard errors in parentheses.
- Reported coefficients:
  - Public Debt: 0.019*** (0.002) in column (1)
  - Public Debt × Post-GFC: -0.005* (0.003) in column (1)
  - External Debt: 0.004*** (0.000) in column (2)
  - External Debt × Post-GFC: -0.001** (0.001) in column (2)
  - Country Fixed Effects: Yes (both columns)
  - Observations: 516 (public debt), 528 (external debt)
  - R2: 0.671 (public debt), 0.646 (external debt)
- Note: EMBI = emerging markets bond index; GFC = global financial crisis.

### Fiscal Consolidation Speed — Simulation of Half-Life to Reduce Debt Shock
- Objective: compute time implied by estimated fiscal reaction functions required to reduce half of the increase in public debt to GDP after an unexpected shock to the stock of public debt.
- Simulation setup:
  - Assumes stable initial public debt ratio d̅ with a shock in a single year.
  - Two interest-growth differential scenarios: low (r−g = zero percent) and high (r−g = 2 percent).
  - Debt path specified as a difference equation using the estimated fiscal reaction function coefficient:
    - d_t − d̅ = (d_{t−1} − d̅) (1 + (r−g) − β̂)
  - Initial conditions: d_0 = d̅ and d_1 = d_0 + ε, where ε is a positive shock to debt.
  - Implied half-lives do not depend on levels of d̅ or ε.
- The chapter references Figure 2.9.3 in the main text as plotting these illustrative simulations.

### Drivers of Emerging Market Resilience (Online Annex 2.6)
- Regression for predicted cumulative change in log real GDP growth and log CPI 12 months after start of risk-off episode:
  - y_{i,τ+12} − y_{i,τ,−1} = Σ_{k=1}^{K} β_k P_{i,τ−1}^k + Σ_{j=1}^{J} γ_j ΔX_{i,τ,−j} + α_τ + ε_{i,τ}
  - Where P_{i,τ−1}^k is a set of seven pre-determined policy variables: anchoring of inflation expectations, reserve adequacy, FX mismatches, macroprudential regulations, external debt burden, cyclically-adjusted fiscal balance, and net tightening of capital flow management measures.
  - ΔX_{i,τ,−j} denotes six lags of changes in log real GDP and log CPI to control for pre-trends.
  - α_τ denotes episode fixed effects.
- Policy Determinants (Online Annex Table 2.6.1) — selected coefficients (dependent variables: Real GDP Growth and Consumer Price Inflation; policy variables entered one by one and expressed in standard deviations):
  - Anchoring of Inflation Expectations:
    - Real GDP Growth column (1): 0.001 (0.005)
    - Consumer Price Inflation column (2): -0.023** (0.009)
  - Reserve Adequacy:
    - Real GDP Growth column (3): 0.006** (0.003)
    - Consumer Price Inflation column (4): -0.005 (0.004)
  - FX Mismatches:
    - Real GDP Growth column (5): -0.009*** (0.003)
    - Consumer Price Inflation column (6): 0.009** (0.005)
  - Macroprudential Policy:
    - Real GDP Growth column (7): 0.005* (0.003)
    - Consumer Price Inflation column (8): -0.005** (0.003)
  - External Debt Burden:
    - Real GDP Growth column (9): -0.008*** (0.002)
    - Consumer Price Inflation column (10): 0.011*** (0.004)
  - Cyclically-adjusted Balance:
    - Real GDP Growth column (11): 0.004 (0.002)
    - Consumer Price Inflation column (12): -0.006* (0.003)
  - Episode Fixed Effects: Yes (all columns)
  - Observations by column: 273, 347, 270, 378, 378, 344, 273, 347, 270, 378, 378, 344
  - R2 by column (selected): 0.377, 0.403, 0.404, 0.362, 0.380, 0.360, 0.632, 0.64, 0.658, 0.56, 0.568, 0.603
  - Note: CPI = consumer price index; FX = foreign exchange. Significance: *** p<0.01; ** p<0.05; * p<0.1.
- External Conditions (Online Annex Table 2.6.2) — regressions of common component of 12-month change in log real GDP or log CPI on external condition proxies:
  - Real GDP Growth, AEs: 0.010 (0.006) in column (1) for Real GDP Growth; 0.020** (0.007) in column (2) for Consumer Price Inflation
  - Commodity Terms of Trade: 0.016* (0.009) for Real GDP Growth; 0.021* (0.010) for Consumer Price Inflation
  - Financial Conditions Impulse on Growth (FCI-G): -0.002 (0.009) for Real GDP Growth; 0.033*** (0.010) for Consumer Price Inflation
  - Observations: 16 (both columns)
  - R2: 0.488 (Real GDP Growth), 0.539 (Consumer Price Inflation)
- Contributions computation:
  - Contributions of policy frameworks: Σ β̂_k ΔP̅_{i,k}^{τ} over significant policy variables, where ΔP̅_{i,τ−1}^k is the mean of the median value of policy variable k during post-GFC episodes relative to pre-GFC.
  - Contributions of external conditions: y̅_τ = α + Σ_{m=1}^{M} β̂_m G_τ^m + ε_τ; contribution = Σ β̂_m G_τ^m, with G_τ^m being real GDP growth in advanced economies relative to trend, commodity terms-of-trade shocks, and US financial conditions. All variables expressed as cumulative changes during the 12 months following the start of the risk-off episode.
- Robustness:
  - Robustness checks exclude individual risk-off episodes (including taper tantrum starting June 2013) one at a time; missing observations for policy variables imputed using median values during a specific risk-off episode.
  - Online Annex Figure 2.6.1 reports distributions of estimated contributions of policy frameworks and external conditions to real GDP growth and inflation for the median emerging market, showing medians and interquartile ranges when excluding one episode at a time.

### The Q-IPF Model (Online Annex 2.7)
- Purpose: Quantitative Integrated Policy Framework (Q-IPF) extends the two-country New Keynesian workhorse for small open economies to jointly consider monetary, FX, and macroprudential policies, accounting for real and nominal rigidities and financial market imperfections that generate inefficient risk premia fluctuations, persistent exchange rate pass-through, and sudden stops due to occasionally binding external debt limits.
- Financial market structure (summary):
  - Financiers intermediate cross-border borrowing and currency conversion, actively trade FX, absorb currency risk, and hold symmetric nominal positions in domestic and foreign bonds (yields I and I*). Financiers have finite risk-bearing capacity; increased bond holdings imply depreciation and a UIP risk premium governed by FX market depth. Home households own a fraction of financiers.
  - Banks transfer funds in domestic currency between financiers and households, face a borrowing constraint proportional to nominal economy size; when debt limit is reached, a positive credit spread arises (domestic bond yield I over domestic risk-free interest I_b). Banks are entirely held by domestic households.
  - Portfolio investors take positions in home-currency bonds financed by issuing foreign-currency bonds. Domestic households own a fraction of portfolio investors.
- Households and firms:
  - Households trade bonds in local currency paying risk-free nominal interest I_b, supply differentiated labor with Dixit-Stiglitz aggregation, set wages in staggered Calvo style, and face mechanism for endogenous inflation expectations anchoring (degree of anchoring depends inversely on deviation of inflation from target).
  - Firms: intermediate goods produced under monopolistic competition with Cobb-Douglas production; prices set in staggered Calvo contracts; non-optimizing firms index to past prices and inflation.
- Monetary and fiscal policy:
  - Monetary policy: standard Taylor rule for short-term interest rate I; central bank minimizes deviations of inflation from target and output from potential; central bank can engage in sterilized FX interventions, taking non-negative positions in foreign bonds and issuing domestic sterilization bonds; financial outcomes of FX interventions are borne by fiscal authority.
  - Fiscal policy: aggregate public consumption exogenous, government keeps budget balanced using lump sum taxes on households.
- Market clearing conditions: domestically produced goods, domestic labor and capital markets, financial market positions where financier holdings match net demand for foreign currency and sterilization bonds issued by monetary authority when engaged in FX interventions.
- FX mismatches:
  - Fraction of financiers owned by home households and fraction of portfolio investors owned by home households jointly determine degree of FX mismatches and currency denomination of net foreign liabilities.
- Calibration:
  - Calibrated to two small open emerging markets: average emerging market in pre-GFC (1997–2009) and post-GFC (2010–24) periods.
  - Post-GFC calibration differences: better anchored inflation expectations; smaller balance sheet mismatches (higher net foreign assets and larger share of external liabilities denominated in domestic currency). All other parameters calibrated based on Adrian and others (2021).
- Exchange rate pass-through to inflation:
  - Estimated as cumulative percentage change in CPI in response to a one percentage point depreciation of the nominal exchange rate vis-à-vis the USD, with country fixed effects, time fixed effects, and lagged controls.
  - Exchange rate depreciation instrumented using change between the day before and the day after scheduled US Federal Reserve Board monetary policy announcements.
  - Pre-GFC period defined as 1997–2009 and post-GFC period as 2010–24; shaded areas in calibration figures denote 90 percent confidence intervals and percentile ranges as specified.

*Source: Online annexes and tables from the IMF World Economic Outlook chapter materials (ch2onlineannexes).*

### Annex Figure 2.7.2.1 and Online Annex Table 2.7.1. The parameters governing price and wage

### Annex Figure 2.7.2.1 and Online Annex Table 2.7.1. The parameters governing price and wage

### Price and wage formation calibration
- Parameters governing price and wage formation in the model are set to match higher exchange rate pass-through in the pre-GFC emerging market relative to the post-GFC emerging market.

### Balance sheet mismatches and related calibration choices
- The share of portfolio investors assets to GDP and the share of financiers and portfolio investors owned by domestic households are set to be lower in the post-GFC period.
- These adjustments are made to match empirically observed increases in net foreign assets and the share of external liability in domestic currency (referenced as Online Annex Figure 2.7.2.2 and Online Annex Figure 2.7.2.3 in the source).
- Ideally, the model could also be calibrated to match changes in the degree of FX market risk-bearing capacity, but data limitation makes this calibration challenging.
- Intuition: increased FX market depth would make responses to a given capital outflow weaker for the average emerging market in the post-GFC period.

### Exchange rate pass-through to inflation (Online Annex Table 2.7.1)
- Note: The table reports the exchange rate pass-through estimated as the cumulative percentage change in the consumer price index in response to a one percentage point depreciation of the nominal exchange rate vis-à-vis the USD, with country fixed effects, time fixed effects, and lagged controls. The exchange rate depreciation is instrumented using the change between the day before and the day after scheduled monetary policy announcements by the US Federal Reserve Board. Standard errors are clustered by time and reported in parentheses. FX = foreign exchange; GFC = global financial crisis.
- Estimated coefficients (cumulative CPI response to a one percentage point nominal exchange rate depreciation vis-à-vis the USD):
  - t + 6:
    - FX Depreciation: 0.332*** (0.121)
    - FX Depreciation x Post-GFC: -0.319** (0.144)
  - t + 12:
    - FX Depreciation: 0.704*** (0.206)
    - FX Depreciation x Post-GFC: -0.529** (0.258)
  - t + 24:
    - FX Depreciation: 1.358*** (0.403)
    - FX Depreciation x Post-GFC: -1.189** (0.510)
  - t + 36:
    - FX Depreciation: 1.767*** (0.608)
    - FX Depreciation x Post-GFC: -1.956** (0.787)
- Regression details:
  - Country Fixed Effects: Yes (all specifications)
  - Time Fixed Effects: Yes (all specifications)
  - Observations: 5,291; 5,171; 4,931; 4,691 (corresponding to t + 6, t + 12, t + 24, t + 36)
  - Adjusted R2: 0.662; 0.672; 0.682; 0.720 (corresponding to t + 6, t + 12, t + 24, t + 36)
- Significance notation:
  - *** p<0.01; ** p<0.05; * p<0.1

### Model simulations, policy tradeoffs, and sudden stops
- The analysis conducts model simulations to illustrate policy tradeoffs and the appropriate mix and timing of policy responses in emerging markets with varying quality of policy frameworks.
- The analysis estimates the frequency of sudden stops with reduced financial frictions using stochastic simulations.
- Calibration approach:
  - The model is first calibrated to match selected second moments of a median emerging market, incorporating both risk-off shocks and conventional demand and supply shocks.
  - These shocks are then used to simulate the model under two alternative calibrations representing emerging markets with weaker and stronger policy frameworks.
  - The external borrowing constraint in the pre-GFC calibration is set such that the economy remains in the constrained regime approximately 3 percent of the time.
- Simulations are run by keeping the distance between the steady-state external debt and its limit constant across simulations.

### Key macroeconomic statistics in Emerging Markets and Advanced Economies (Online Annex Table 2.8.1)
- Note: The table reports key statistics from Aguiar and Gopinath (2007) for the emerging markets in the sample pre-GFC and post-GFC and compares these statistics to advanced economies. The pre-GFC period is 1997–2009, and the post-GFC period is 2010–24. AEs = advanced economies; C = real consumption; EMs = emerging markets; GFC = global financial crisis; SD = standard deviation; TB = trade balance to GDP ratio; Y = real GDP; ρ = correlation.
- Reported statistics (columns correspond to Pre-GFC, Post-GFC, AEs):
  - SD(ln Yt): 2.75, 2.01, 1.69
  - SD(Δln Yt): 1.87, 1.37, 1.04
  - SD(ln Ct): 2.71, 2.53, 1.52
  - SD(TBt): 4.33, 4.27, 3.54
  - ρ(ln Yt, ln Yt-1): 0.74, 0.76, 0.80
  - ρ(Δln Yt, Δln Yt-1): 0.08, -0.08, 0.06
  - SD(ln Ct) / SD(ln Yt): 1.01, 1.25, 0.89
  - ρ(TBt, ln Yt): -0.29, -0.17, -0.06
- Summary interpretation from the source:
  - Prior to the GFC, emerging markets exhibited higher income volatility, more volatile consumption relative to output, and a more counter‐cyclical trade balance than advanced economies.
  - In the post-GFC sample, income shocks in emerging markets have become less pronounced and the trade balance has become less procyclical.
  - Despite this convergence, differences with advanced economies remain.
  - Pre-GFC period defined as 1997–2009; post-GFC period defined as 2010–24.

*Source: Annex Figure 2.7.2.1 and Online Annex Table 2.7.1 in ch2onlineannexes (International Monetary Fund, World Economic Outlook, October 2025).*

---


_Source: https://www.imf.org/-/media/files/publications/weo/2025/october/english/ch2onlineannexes.pdf_