## ch3onlineannex

## Source details

**Canonical URL:** [ch3onlineannex](https://www.imf.org/-/media/files/publications/reo/whd/2025/october/english/ch3onlineannex.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/reo/whd/2025/october/english/ch3onlineannex.pdf.md)
- [Structured JSON version](/-/media/files/publications/reo/whd/2025/october/english/ch3onlineannex.pdf.json)

---

### Chief Executive and Board: rules governing appointment, dismissal, term length, and reappointment
- Policy and institutional features covered:
  - Appointment, dismissal, term length, and reappointment rules for the governor and board members.
  - Key questions: How is the central bank governor appointed? Are there restrictions on dismissal? What is the governor’s term of office?
- Related CBI dimensions referenced in analysis:
  - Policy Formulation: autonomy in setting monetary policy, exchange rate policy, and supervisory functions; exclusive authority to set monetary policy; responsibility for exchange rate policy.
  - Central Bank Objectives: whether price stability is designated as the primary objective and whether objectives are clearly specified in law.
  - Lending to Government: legal limits or prohibitions on direct and indirect central bank lending to the government.
  - Financial Independence: autonomy in budgeting, income retention, and control over financial resources; ability to retain earnings.
  - Transparency and Accountability: reporting requirements, audit processes, publication of data, and communication with the public or parliament.
- Source attribution within annex: Romelli (2024).

### Methodology: panel regressions, local projections, and quantile regressions
- Baseline fixed-effects panel model (Equation (1)):
  - Dependent variable: transformed annual percentage change in the consumer price index (CPI) in country i at time t.
  - Inflation transformation used (rescaled formula) to reduce influence of outliers: 휋휋푖푖,푡푡 = (transformation described in text).
  - Key explanatory variable: CBI index by Romelli (2024), measured annually for each country.
  - Lagged controls (푋푋푖푖,푡푡−1): lagged output gap; lagged inflation.
- Robustness and alternative measures:
  - Alternative legal CBI measures: Cukierman, Webb, and Neyapti (1992) and Garriga (2025).
  - Additional controls: exchange rate regime (Ilzetzki, Reinhart, and Rogoff (2019) classification); indicators of systemic banking crises (Laeven and Valencia (2020)); international inflation proxied by US CPI inflation; measures of structural reforms.
- Local projections (Jordà (2005)) for cumulative inflation over ten-year horizon (ℎ = 1, ..., 10):
  - Equation (2) specification provided to trace dynamic response of inflation to changes in CBI and to assess regional heterogeneity (Latin America and the Caribbean highlighted).
  - Cumulative inflation definition: (CPIi,t+h / CPIi,t) − 1 ≈ log(CPIi,t+h) − log(CPIi,t).
- Panel quantile regression (Equation (3)):
  - Purpose: assess whether higher CBI is associated with lower inflation in the upper tail of the distribution and test which CBI subcomponents (policy formulation, central bank objectives, etc.) moderate extreme inflation outcomes.

### Response to monetary policy shocks: data, sample, and local projections
- Monetary policy shocks definition: Checo, Grigoli, and Sandri (2024) — difference between realized policy rate and average forecast from market participants immediately before the meeting.
- Final sample (monetary policy shocks): Brazil, Chile, Colombia, Egypt, Hungary, India, Indonesia, Malaysia, Mexico, Nigeria, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand, Türkiye.
- Monthly additional variables: CPI; exchange rate (local currency per US dollar); long-term yields (from WEO); 12-month-ahead inflation expectations (Consensus Economics); debt and nominal GDP from WEO for debt-to-GDP.
- Local projections cumulative impulse response regression:
  - Interaction term: dummy = 1 if debt-to-GDP > 80th percentile for each country (country-specific high debt indicator).
  - Outcome variables: inflation; 12-month-ahead inflation expectations; log exchange rates; 10-year nominal yields; credit as share of GDP; unemployment rate.
  - Chapter 3, Figure 3.9 shows responses at 18-month horizon for four alternative outcome variables.
- Interpretation of low vs high debt interactions (Online Annex Figure 3.1):
  - Panel A (Indicator = 0, low debt): after a monetary policy shock the exchange rate appreciates, the inflation rate decreases, long-term yields drop, and inflation expectations decrease.
  - Panel B (Indicator = 1, high debt): after a monetary policy shock the exchange rate does not appreciate, the inflation rate does not decrease, long-term yields do not drop, and inflation expectations decrease.
- Robustness checks:
  - Inclusion of time and country fixed effects; lagged changes of the dependent variable; percentiles of debt at 70 percent and 85 percent.
  - Results qualitatively similar for horizons: 12 months, 24 months, 36 months.
  - Online Annex Figure 3.1 contains full impulse responses for a 36-month horizon.
- Discussion takeaway:
  - Explanations for inflation increases after a monetary policy shock (“price puzzle”) include information effects and fiscal-policy-related channels (rising risk premiums from higher deficits, exchange rate depreciation, higher inflation expectations, or rising total debt increasing nominal demand via wealth effects).
  - Key takeaway: evidence points to the importance of fiscal backing of monetary policy to ensure monetary policy objectives can be attained.

### Selected Online Annex figures (high-level)
- Online Annex Figure 3.1. Debt Levels and the Response to Monetary Policy Shocks:
  - Panels A (Low Debt Levels) and B (High Debt Levels) show cumulative effects on: 1. Inflation (Percent); 2. Inflation Expectations (Percent); 3. Exchange Rate (Local currency per US dollar; percent); 4. Ten-year Yields (Percent).
  - Note: interaction = indicator for debt higher than 80th percentile. Exchange rate units are local currency per US dollar (decrease is an appreciation).
- Online Annex Figure 3.2. Impact of Fiscal Consolidations on Two-year Ahead Inflation Expectations as a Function of Debt-to-GDP:
  - Dots show cumulative responses over a three-year period following a 1-percent-of-GDP shock. Bars show 90th percentile.

### Response to fiscal policy shocks: data, sample, methodology, and additional results
- Sample: same EMDE sample as monetary policy exercise (Brazil, Chile, Colombia, Egypt, Hungary, India, Indonesia, Malaysia, Mexico, Nigeria, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand, Türkiye).
- Fiscal deficit shocks: WEO forecast errors on structural primary balance (April vintages). Shock for year t = forecasted structural primary balance for year t in April − realized outturn for structural primary balance.
- Annual data: real GDP and headline inflation (WEO); 12-month-ahead inflation expectations (Consensus Economics).
- Methodology: local projections (Jordà (2005)) for cumulative impulse response:
  - Controls include current and lagged real GDP growth and debt-to-GDP.
  - Identifying assumption: projection error uncorrelated with economic conditions.
- Additional results:
  - Arizala and others (forthcoming) compute responses of inflation and inflation expectations to changes in cyclically adjusted primary balance for different debt levels (Online Annex Figure 3.2).
  - Chapter 3, Figure 3.11 depicts impulse responses of inflation at one- and two-year horizons with 95 percent confidence intervals.

### Interest rate rules: Taylor-type regressions and debt interactions
- Sample: same EMDE sample (Brazil, Chile, Colombia, Egypt, Hungary, India, Indonesia, Malaysia, Mexico, Nigeria, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand, Türkiye).
- Data: policy rates from Haver; headline inflation, output gap, and debt-to-GDP from WEO (quarterly frequency); exchange rate regime variable included (Ilzetzki, Reinhart, and Rogoff (2019) classification).
- Regression specification (Taylor rule):
  - 훾푖: country fixed effect; 훼퐼푖,푡−1 captures smoothing in policy rate choices.
  - Indicator if debt > 80th percentile enters via 퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼 푖,푡.
  - Responses of policy rates to inflation reported as: 훽1 / (1 − 훼) and (훽1 + 훽2) / (1 − 훼) (depicted in Chapter 3, Figure 3.9).

### Fiscal impact of monetary policy: debt-stabilizing primary balance and illustrative numbers
- Debt-stabilizing primary balance formula:
  - 푝푝푏푏∗ = 푑푑∗ × (퐼퐼 − 휋휋 − 푔푔).
  - 퐼 = 훼 �1 + 퐼푓� + (1 − 훼) �1 + 퐼푑� − 1 (weighted average of local and foreign currency debt).
- Illustrative results:
  - With medium run real interest rates around 3 percent, the debt-stabilizing primary balance is approximately 0.6 percent of GDP (orange square in Online Annex Figure 3.3).
  - Blue dots illustrate effect of a 100-basis point increase in nominal interest rates on total interest payments, holding growth and 2024 debt fixed, resulting in higher debt-stabilizing primary balances.
  - As interest rates rise, the gap between necessary debt-stabilizing primary balances and current primary balances widens significantly.
- Discussion notes:
  - The measure is approximate: marginal rates differ from average rates and long debt maturities imply slow feed-through from marginal to average rates.
  - Indexed debt, floating-rate debt, and short maturities amplify these forces.
- Online Annex Figure 3.3:
  - LA7: Debt-stabilizing Primary Balance, Gross Public Debt, and Nominal Interest Rate.
  - LA7 = Latin America 7 (Brazil, Chile, Colombia, Mexico, Paraguay, Peru, Uruguay).
  - Orange dot: debt-stabilizing primary balance given 2024 debt-to-GDP and medium-run growth and average interest payments minus inflation.
  - Blue dots: computed by increasing or decreasing average nominal rate of local currency debt by 100 basis points, holding growth, debt-to-GDP, and payments of foreign currency debt fixed.

### Monetary–fiscal interactions: framework and channels
- Channels and mechanisms (Online Annex Figure 3.4):
  - Monetary policy shocks affect macroeconomy and inflation via exchange rates, expectations, term structure of interest rates, and credit and asset prices.
  - Debt levels increase risk and liquidity premiums, amplify risk of elevated and volatile inflation, and influence inflation expectations and exchange rate risk premium.
  - Fiscal policy affects inflation through the Keynesian aggregate demand channel and via the government budget constraint that mediates monetary–fiscal interaction (fiscal policy influences debt, which affects bond spreads, inflation expectations, and exchange rate risk premium).
  - High debt levels amplify aggregate demand via wealth effects and heighten risk of central bank accommodating fiscal needs (fiscal dominance risks).
- Notation in figure includes: 퐼 (interest rate); Δ퐼 (depreciation); x (output gap); E(휋′|H) (expected inflation); R C(B), L C (risk and liquidity premium); ΔM (change in Monetary Base); ΔB (change in debt); B (debt).

### Time-varying response to monetary policy shocks
- Methodology: follows Online Annex 2.3 of IMF (2024) extended to larger sample ending in 2024:Q4; quarterly data starting 1995:Q1 through 2024:Q4.
- Countries shown in Online Annex Figure 3.5: Brazil, Chile, Colombia, Mexico, Peru.
- Model estimated with Bayesian techniques following Del Negro and Primicieri (2015) and Gambetti and Musso (2017).
- Online Annex Figure 3.5 shows effect on consumer price index of a 100 basis points monetary policy tightening (increase in policy rate).

### Additional evidence and the quantitative magnitude of CBI effects
- Main finding on central bank independence (CBI):
  - A 0.25 increase in central bank independence is associated with a 1 percentage point reduction in inflation.
  - This magnitude is comparable to:
    - the current gap between the 25th and 75th percentiles of the CBI distribution, or
    - the gap between Latin America and the Caribbean (LAC) countries and the Organisation for Economic Co-operation and Development (OECD) average.
  - The magnitude also aligns with the size of reforms implemented in many LAC countries during the 1990s.
- Robustness and controls:
  - Relationship robust to inclusion of exchange rate regime (Ilzetzki, Reinhart, and Rogoff 2019); indicators of systemic banking crises Laeven and Valencia (2020); international inflation proxied by US CPI inflation; measures of structural reforms.
  - Results come from baseline regression of inflation on CBI indices controlling for one-year lag of inflation, the output gap, country and time fixed effects; dependent variable is the re-scaled CPI; standard errors clustered at the country level.
- Heterogeneous effects and tail-risk mitigation (quantile regression):
  - Overall CBI index has a stronger disinflationary impact at higher inflation levels—particularly in LAC, where inflation has historically been elevated.
  - Within LAC, CBI sub-indices' effects rank as:
    - Policy dimension: most pronounced effect;
    - Objective dimension: second most pronounced effect.
  - Implication: Policy and Objective represent foundational aspects of CBI necessary for other dimensions to be effective.
- Figures and notes referenced:
  - Online Annex Box Figure 3.1.1: Beta Coefficients of CBI Index on Annual Inflation (Percentage points; 95 percent confidence intervals).
  - Online Annex Box Figure 3.1.2: LAC: CBI Sub-Index Across the Inflation Distribution (Δπ per CBI unit; re-scaled CPI; CBI index: 0–1).
- Authorship: the author of this box is Agnese Carella.

*Source: ch3onlineannex — Online Annex text and figures as provided in the source content.*

### 1. Chief Executive and Board Rules governing the appointment, dismissal,

### 1. Chief Executive and Board Rules governing the appointment, dismissal, term length, and reappointment of the governor and board members.

### Policy and institutional features covered
- Appointment, dismissal, term length, and reappointment rules for the governor and board members.
- Questions addressed:
  - How is the central bank governor appointed?
  - Are there restrictions on dismissal?
  - What is the governor’s term of office?

### Related CBI (central bank independence) dimensions referenced in analysis
- Policy Formulation: autonomy in setting monetary policy, exchange rate policy, and supervisory functions; exclusive authority to set monetary policy; responsibility for exchange rate policy.
- Central Bank Objectives: whether price stability is designated as the primary objective and whether objectives are clearly specified in law.
- Lending to Government: legal limits or prohibitions on direct and indirect central bank lending to the government.
- Financial Independence: autonomy in budgeting, income retention, and control over financial resources; ability to retain earnings.
- Transparency and Accountability: reporting requirements, audit processes, publication of data, and communication with the public or parliament.

*Source: Romelli (2024).*

---

### 3.3. Methodology (Panel regressions and robustness checks)

### Baseline specification (fixed-effects panel model)
- Equation (1): 휋휋푖푖,푡푡 = 훼훼푖푖 + 훽훽훽훽훽훽 퐼퐼푖푖,푡푡 + 훾훾푋푋푖푖,푡푡−1 + 휀휀푖푖,푡푡
- Dependent variable: transformed annual percentage change in the consumer price index (CPI) in country i at time t.
- Inflation transformation used (rescaled formula) to reduce influence of outliers:
  - 휋휋푖푖,푡푡 = (I I I I I I I I I I I I I I I I I I  I I  퐼퐼푖푖,푡푡) / (1 + I I I I I I I I I I I I I I I I I I  퐼퐼푖푖,푡푡)  (transformation described; commonly used in the literature).
- Key explanatory variable: the CBI index by Romelli (2024), measured annually for each country.
- Lagged controls (푋푋푖푖,푡푡−1) in baseline:
  - Lagged output gap.
  - Lagged inflation.

### Robustness and alternative measures
- Alternative legal CBI measures: Cukierman, Webb, and Neyapti (1992) and Garriga (2025).
- Additional controls for robustness:
  - Exchange rate regime (Ilzetzki, Reinhart, and Rogoff (2019) classification).
  - Indicators of systemic banking crises (Laeven and Valencia (2020)).
  - International inflation proxied by US CPI inflation.
  - Measures of structural reforms.
- Long-run dynamics estimated using local projections (Jordà (2005)) to trace responses over ten-year horizon (ℎ = 1, ..., 10).

### Local projections for cumulative inflation
- Equation (2): 푦푦푖푖,푡푡,푡푡+ℎ = 훼훼푖푖 + 훿훿푡푡 + 훽훽1 훽훽훽훽퐼퐼푖푖,푡푡 + 훽훽2 푋푋푖푖푡푡−1 + 휀휀푖푖,푡푡
- 푦푦푖푖,푡푡,푡푡+ℎ = (CPI푖푖,푡푡+ℎ / CPI푖푖,푡푡) − 1 ≈ log(CPI푖푖,푡푡+ℎ) − log(CPI푖푖,푡푡)
- Purpose: trace dynamic response of inflation to changes in CBI over ten years and assess regional heterogeneity (Latin America and the Caribbean highlighted).

### Panel quantile regression
- Equation (3): 휋휋푖푖푡푡푞 = 훼훼푖푖푡푡푞 + 훽훽푞 훽훽훽훽퐼퐼푖푖,푡푡 + 휀휀푖푖푡푡푞
- Purpose: assess whether higher CBI is associated with lower inflation in the upper tail of the distribution and to test which CBI subcomponents (policy formulation, central bank objectives, etc.) are influential in moderating extreme inflation outcomes.

---

### 3.4. The Response to Monetary Policy Shocks

### Data and sample
- Monetary policy shocks: Checo, Grigoli, and Sandri (2024) — difference between realized policy rate and average forecast from market participants immediately before the meeting.
- Sample inclusion criteria:
  - Short-term policy rate movements can be reliably tracked.
  - Use of conventional monetary policy tools with well-defined policy rate instruments.
  - Sufficiently long forecast time series to construct monetary policy surprises.
- Final sample: Brazil, Chile, Colombia, Egypt, Hungary, India, Indonesia, Malaysia, Mexico, Nigeria, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand, Türkiye.
- Monthly additional variables: CPI, exchange rate (local currency per US dollar), long-term yields (from WEO), 12-month-ahead inflation expectations (Consensus Economics). Debt and nominal GDP from WEO for debt-to-GDP.

### Methodology (local projections)
- Cumulative impulse response regression:
  - Y푐,푡+ℎ − Y푐,푡 = 훼푐ℎ + 훿푡ℎ + 훽퐼ℎ 퐼푡 + 훾퐼ℎ (Indicator for high debt) × 퐼푡 + u푖푡
- Interaction term: dummy that equals 1 if debt-to-GDP ratio is higher than the 80th percentile for each country (country-specific high debt indicator).
- Outcome variables: inflation, 12-month-ahead inflation expectations, log exchange rates, 10-year nominal yields, credit as share of GDP, unemployment rate.
- Chapter 3, Figure 3.9 shows responses at 18-month horizon for four alternative outcome variables.

### Interpretation of low vs high debt interactions
- Online Annex Figure 3.1 Panel A (Indicator = 0, low debt):
  - Upon a monetary policy shock the exchange rate appreciates, the inflation rate decreases, long-term yields drop, and inflation expectations decrease.
- Panel B (Indicator = 1, high debt):
  - Upon a monetary policy shock the exchange rate does not appreciate, the inflation rate does not decrease, long-term yields do not drop, and inflation expectations decrease.

### Robustness checks
- Inclusion of time and country fixed effects; lagged changes of the dependent variable; percentiles of debt at 70 percent and 85 percent.
- Results qualitatively similar for horizons: 12 months, 24 months, 36 months.
- Online Annex Figure 3.1 contains full impulse responses for a 36-month horizon.

### Discussion takeaway
- Literature finds inflation increases after a monetary policy shock (“price puzzle”) possibly due to:
  - Information effect (policymakers respond to inflation signals unavailable to econometrician).
  - Fiscal-policy-related channels: rising risk premiums from higher deficits leading to exchange rate depreciation, higher inflation expectations, or rising total debt increasing nominal demand via wealth effects.
- Key takeaway: evidence points to the importance of fiscal backing of monetary policy to ensure monetary policy objectives can be attained.

---

### Online Annex Figures and Notes (selected)

### Online Annex Figure 3.1. Debt Levels and the Response to Monetary Policy Shocks
- Panels A (Low Debt Levels) and B (High Debt Levels) show cumulative effects on:
  1. Inflation (Percent)
  2. Inflation Expectations (Percent)
  3. Exchange Rate (Local currency per US dollar; percent)
  4. Ten-year Yields (Percent)
- Note: Interaction is indicator for debt higher than 80th percentile. Exchange rate units are local currency per US dollar (decrease is an appreciation).

### Online Annex Figure 3.2. Impact of Fiscal Consolidations on Two-year Ahead Inflation Expectations as a Function of Debt-to-GDP
- Sources: Arizala and others (forthcoming); Consensus Economics; IMF, World Economic Outlook database; IMF staff calculations.
- Note: Dots show cumulative responses over a three-year period following a 1-percent-of-GDP shock. Bars show 90th percentile.

---

### 3.5. The Response to Fiscal Policy Shocks

### Data and sample
- Same EMDE sample as monetary policy exercise: Brazil, Chile, Colombia, Egypt, Hungary, India, Indonesia, Malaysia, Mexico, Nigeria, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand, Türkiye.
- Fiscal deficit shocks: WEO forecast errors on structural primary balance (April vintages). Shock for year t = forecasted structural primary balance for year t in April − realized outturn for structural primary balance.
- Annual data: real GDP and headline inflation (WEO); 12-month-ahead inflation expectations (Consensus Economics).

### Methodology
- Local projections (Jordà (2005)) for cumulative impulse response:
  - Y푐,푡+ℎ − Y푐,푡 = 훼푐ℎ + 훿푡ℎ + 훽퐼ℎ F C푡 + Σ훾ℎ훽퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼  푠푖푡−푘 + u푖푡
- Controls include current and lagged real GDP growth and debt-to-GDP.
- Identifying assumption: projection error uncorrelated with economic conditions.

### Additional results
- Arizala and others (forthcoming) compute responses of inflation and inflation expectations to changes in cyclically adjusted primary balance for different debt levels (Online Annex Figure 3.2).
- Chapter 3, Figure 3.11 depicts impulse responses of inflation at one- and two-year horizons with 95 percent confidence intervals.

---

### 3.6. Interest Rate Rules

### Data and sample
- Same EMDE sample (Brazil, Chile, Colombia, Egypt, Hungary, India, Indonesia, Malaysia, Mexico, Nigeria, Peru, Philippines, Poland, Romania, Russia, South Africa, Thailand, Türkiye).
- Policy rates from Haver; headline inflation, output gap, and debt-to-GDP from WEO (quarterly frequency).
- Exchange rate regime variable included (Ilzetzki, Reinhart, and Rogoff (2019) classification).

### Methodology (Taylor rule specification)
- Regression: 퐼퐼푖,푡 = 훾푖 + 훼퐼푖,푡−1 + 훽1 휋푖,푡 + 훽2 휋푖,푡 퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼 푖,푡 + 훾푖 x푖,푡 + 휖푖,푡
- 훾푖: country fixed effect.
- 퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼퐼 푖,푡: indicator if debt > 80th percentile.
- 훼퐼푖,푡−1 captures smoothing in policy rate choices.
- Responses of policy rates to inflation: 훽1 / (1 − 훼) and (훽1 + 훽2) / (1 − 훼) (depicted in Chapter 3, Figure 3.9).

---

### 3.7. Fiscal Impact of Monetary Policy

### Theory and key formulas
- Debt-stabilizing primary balance formula:
  - 푝푝푏푏∗ = 푑푑∗ × (퐼퐼 − 휋휋 − 푔푔)
  - 푝푝푏푏∗: constant primary balance that stabilizes debt 푑푑∗ given average nominal interest rate on debt 퐼, inflation 휋, growth g.
  - 퐼 = 훼 �1 + 퐼푓� + (1 − 훼) �1 + 퐼푑� − 1 (weighted average of local and foreign currency debt).

### Results and illustrative numbers
- With medium run real interest rates around 3 percent, the debt-stabilizing primary balance is approximately 0.6 percent of GDP (orange square in Online Annex Figure 3.3).
- Blue dots illustrate effect of a 100-basis point increase in nominal interest rates on total interest payments, holding growth and 2024 debt fixed, resulting in higher debt-stabilizing primary balances.
- As interest rates rise, the gap between necessary debt-stabilizing primary balances and current primary balances widens significantly.

### Discussion
- The measure is approximate: marginal rates differ from average rates and long debt maturities imply slow feed-through from marginal to average rates.
- Indexed debt, floating-rate debt, and short maturities amplify these forces.

### Online Annex Figure 3.3
- LA7: Debt-stabilizing Primary Balance, Gross Public Debt, and Nominal Interest Rate.
- LA7 = Latin America 7 (Brazil, Chile, Colombia, Mexico, Paraguay, Peru, Uruguay).
- Orange dot: debt-stabilizing primary balance given 2024 debt-to-GDP and medium-run growth and average interest payments minus inflation.
- Blue dots: computed by increasing or decreasing average nominal rate of local currency debt by 100 basis points, holding growth, debt-to-GDP, and payments of foreign currency debt fixed.

---

### 3.8. Monetary–Fiscal Interactions: A Framework

### Channels and mechanisms (as in Online Annex Figure 3.4)
- Monetary policy shocks affect macroeconomy and inflation via:
  - Exchange rates (interest rate differentials, expected depreciation, risk premium).
  - Expectations (expected inflation guiding price-setting).
  - Term structure of interest rates (short-term rates and risk premiums).
  - Credit and asset prices.
- Debt levels:
  - Increase risk and liquidity premiums.
  - Amplify risk of elevated and volatile inflation.
  - Influence inflation expectations and exchange rate risk premium.
- Fiscal policy:
  - Standard Keynesian aggregate demand channel through fiscal stance affecting inflation.
  - Government budget constraint mediates monetary–fiscal interaction: fiscal policy influences debt, which affects bond spreads, inflation expectations, and exchange rate risk premium.
- High debt levels:
  - Amplify aggregate demand via wealth effects.
  - Heighten risk of central bank accommodating fiscal needs (fiscal dominance risks).

### Notation in figure
- 퐼: interest rate; Δ퐼: depreciation; x: output gap; E(휋′|H): expected inflation; R C(B), L C: risk and liquidity premium; ΔM: change in Monetary Base; ΔB: change in debt; B: debt.

---

### 3.9. Time-varying Response to Monetary Policy Shocks

### Methodology and sample
- Methodology follows Online Annex 2.3 of IMF (2024) extended to larger sample ending in 2024:Q4.
- Countries included: Brazil, Chile, Colombia, Mexico, Peru (Online Annex Figure 3.5).
- Quarterly data starting 1995:Q1 through 2024:Q4.
- Model estimated with Bayesian techniques following Del Negro and Primicieri (2015) and Gambetti and Musso (2017).

### Online Annex Figure 3.5
- Shows effect on consumer price index of a 100 basis points monetary policy tightening (increase in policy rate).

---

### Online Annex Box 3.1. Additional evidence from Panel and Quantile Regression
- Estimating equation (1) using the Romelli index (Romelli (2024)) and extending analysis with panel quantile regressions to probe CBI subcomponents’ effects across the inflation distribution (details in the Online Annex Box).

*Source: IMF staff calculations; Romelli (2024); Checo, Grigoli, and Sandri (2024); Jordà (2005); Ilzetzki, Reinhart, and Rogoff (2019); Laeven and Valencia (2020); Cukierman, Webb, and Neyapti (1992); Garriga (2025); Arizala and others (forthcoming); Consensus Economics; IMF, World Economic Outlook database.*

### 0.25 increase in central bank independence is

### 0.25 increase in central bank independence is 

### Main finding
- A 0.25 increase in central bank independence (CBI) is associated with a 1 percentage point reduction in inflation.
- This magnitude is comparable to:
  - the current gap between the 25th and 75th percentiles of the CBI distribution, or
  - the gap between Latin America and the Caribbean (LAC) countries and the Organisation for Economic Co-operation and Development (OECD) average.
- The magnitude also aligns with the size of reforms implemented in many LAC countries during the 1990s.

### Robustness and controls
- The relationship is robust to inclusion of additional controls, including:
  - the country’s exchange rate regime (classified according to Ilzetzki, Reinhart, and Rogoff 2019),
  - indicators of systemic banking crises Laeven and Valencia (2020),
  - international inflation, proxied by US consumer price index inflation,
  - measures of structural reforms.
- Results are from a baseline regression of inflation on CBI indices controlling for:
  - one-year lag of inflation,
  - the output gap,
  - country and time fixed effects.
- The dependent variable is the re-scaled CPI.
- Standard errors are clustered at the country level.

### Heterogeneous effects and tail-risk mitigation (quantile regression)
- Quantile regression captures heterogeneous effects of CBI across the inflation distribution and assesses its role in mitigating tail risks.
- The overall CBI index has a stronger disinflationary impact at higher inflation levels—particularly in LAC, where inflation has historically been elevated.
- Within LAC, the CBI sub-indices' effects rank as:
  - Policy dimension: most pronounced effect,
  - Objective dimension: second most pronounced effect,
  - (Implication: Policy and Objective represent foundational aspects of CBI necessary for other dimensions to be effective.)

### Figures and methodological notes
- Online Annex Box Figure 3.1.1: Beta Coefficients of CBI Index on Annual Inflation (Percentage points; 95 percent confidence intervals).
  - Sources: Cukierman, Web, and Neyapti (1992); Romelli (2024); Garriga (2025); and IMF staff calculations.
  - Note: Results from baseline regression of inflation on central bank independence indices controlling for one-year lag of inflation and the output gap, country and time fixed effects. The dependent variable is the re-scaled CPI. Standards errors are clustered at the country level. CBI = central bank independence.
- Online Annex Box Figure 3.1.2: LAC: CBI Sub-Index Across the Inflation Distribution (Δπ per CBI unit; re-scaled CPI; CBI index: 0–1).
  - Sources: Romelli (2024); and IMF staff calculations.
  - Note: Quantile regression as in Gelos and others (2022) of transformed inflation on CBI index. The solid line is the point estimate; the dark and light-shaded areas are the 90 and 95 percent confidence bands, respectively. CBI = central bank independence; CPI = consumer price index; EMDE = emerging market and developing economies; LAC = Latin America and the Caribbean.

### Authorship
- The author of this box is Agnese Carella.

*Source: ch3onlineannex — Online Annex Box text from the provided IMF content.*

---


_Source: https://www.imf.org/-/media/files/publications/reo/whd/2025/october/english/ch3onlineannex.pdf_