## Monetary Policy in Disaster-Prone Developing Countries — Working Paper No. WP/2022/067

## Source details

**Canonical URL:** [Monetary Policy in Disaster-Prone Developing Countries — Working Paper No. WP/2022/067](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022067-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022067-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022067-print-pdf.pdf.json)

---

### Methodology (Narrative analysis design and sample)
- Overview:
  - Narrative analysis uses IMF Article IV staff reports prepared in the year of, and one year following, the occurrence of a disaster.
  - Study focuses on disaster-years with annual damages of at least 1 percent of GDP, subject to staff report availability.
  - Narrative records: the nature of mobilized monetary policy tools; whether policy was accommodative, tight, or unchanged; IMF staff and/or Board appraisal of the monetary policy stance; and IMF advice for the near future.
- Sample and data:
  - Final sample: 34 disaster-years, occurring in 16 disaster-prone countries from 1999 to 2017.
  - Damage threshold for inclusion: at least 1 percent of GDP in the disaster-year (subject to report availability).
  - For countries in currency unions, Article IV staff reports for the union’s central bank were cross-referenced.
  - Example country-level facts cited:
    - Damages represent approximately 7% of GDP on average for disaster-prone EMDEs versus 0.5% in their peers less exposed to natural disaster.
    - Extreme events causing damages beyond 200% of GDP are documented.
    - Belize: hurricane Keith (October 2000) and hurricane Iris (October 2001) caused damages “of the tune of 30 percent of GDP each,” and GDP growth in 2001 and 2002 was about 8 percentage points lower than in the pre-shock year.
- Narrative analysis design and coding:
  - Structured questionnaire (Table 2) records answers including Y-N-Mixed-NA options and open responses (e.g., monetary policy tools).
  - Classification of monetary policy independence:
    - “Independent” — full control of monetary policy (own legal tender and policy autonomy).
    - “Not independent” — does not have own legal tender or has a hard peg.
    - “Mixed” — peg or exchange rate anchor but limited capital mobility allows some room for monetary policy.
  - Coding procedure: reading archived Article IV reports and extracting quotes and questionnaire answers; Appendix B documents full narrative analysis with quotes for all disaster-country observations.
- Procedural notes and inclusion rules:
  - Some answers based on consultations occurring two or three years after the disaster due to staff report availability (examples: El Salvador 2011; Micronesia 2015; Solomon Islands 2014; Samoa 2012).
  - For currency-union members (e.g., ECCU), Article IVs for both the country and the union were combined when relevant.
  - Certain distributional figures restrict sample to cases with cumulative loss of at least 5 percent of GDP where noted.

### Link to analytical modeling framework
- Model used: New-Keynesian model with the Taylor projection solution method.
- Extensions considered:
  1. Disaster effects on productivity with permanent and temporary components; disasters affecting export demand.
  2. Small-economy setup along Gali and Monacelli (2005).
  3. Alternative Taylor-type interest rate rules capturing several monetary policy regimes to evaluate welfare outcomes.

### Main empirical findings on macroeconomic impact (narrative results)
- Sample: weather-related natural disasters with damages of at least 1% of GDP → 34 incidents in 16 countries.
- Growth and inflation:
  - In most cases, GDP growth declined in the aftermath of disasters.
  - Inflation often increased, with counterexamples (e.g., Belize): "Real GDP growth declined from 11 percent in 2000 to 5 percent in 2001" and "inflation remained very low at an annual rate of 1¼ percent".
- Monetary policy independence and reserves:
  - Most disasters occurred in countries with at least partial control of their monetary policy; only 12 percent of cases refer to disasters occurred in countries with dollarized economies.
  - If non-pegged countries were considered, the percentage of countries that experienced an impact on their reserves would go down to 35 percent.
- Monetary policy stance after disasters:
  - Stance changed in virtually all cases where there was room for maneuver.
  - When changed:
    - Tightened in almost 56 percent of disasters.
    - Accommodated in the remaining changed cases.
  - Main tool used: interest rate (other tools such as money supply used in several cases).
  - Constraints to changes: hard pegs or dollarized economies.
- IMF staff appraisal and advice:
  - IMF staff and/or directors generally agreed when authorities adopted a tight stance.
  - Directors also agreed with loosening in "about half of instances in which authorities adopted a loose monetary policy stance".
  - No cases where IMF advised to switch from tight to loose in the near future; advice to tighten after a loose stance does occur.

### Robustness and limitations (narrative)
- Narrative results discussed with mission chiefs and desk economists covering about a third of the countries; findings generally align with field experience.
- Noted omissions:
  - Exclusion of the fiscal response or donor support.
  - Exclusion of lower-impact disasters (damages smaller than 1 percent of GDP) due to feasibility and reporting limitations.

### Representative coded excerpts and exact preserved figures (selected)
- Belize, 2000/2001: "Real GDP growth declined from 11 percent in 2000 to 5 percent in 2001"; inflation "remained very low at an annual rate of 1¼ percent"; damages “of the tune of 30 percent of GDP each.”
- Dominica, 2017: "While Erika had caused severe damage, estimated at 96 percent of GDP..."
- Haiti, 2004: "The property damage and the interruption to economic activity are estimated to have totaled 5½ percent of GDP..."
- Haiti, 2004 interest rates: "During August–October 2004, interest rates were reduced to 7.6 percent from 13.6 percent..."
- Jamaica, 2010 policy rate: "the central bank lowered the policy rate to 6.25 percent..."
- Micronesia, 2015 reserves ratio: "The ratio of international reserves to money base was 96 percent at end-2015, compared with the statutory mandate of 60 percent..."
- Moldova, 2007 policy change: "and raised policy interest rates by 2.5 percentage points..."
- Sri Lanka, 2016 policy change: "the Central Bank of Sri Lanka (CBSL) raised the policy rate by 25 basis points..."
- Inflation example: "3.8 percent in 2017 from 0.4 percent in 2016..."

### Transition to quantitative analysis
- Heterogeneity in policy conduct and the multiple stabilization challenges (inflation, business cycle, exchange rate) motivate the quantitative investigation in a SOE New-Keynesian model with stochastic trend growth, disaster shocks, Epstein-Zin preferences, disaster effects on capital and TFP, and alternative Taylor-type rules.

---

### Monetary policy modeling and regimes (baseline and alternatives)
- Generalized feedback rule used:
  - R_t / R = (Π_t / ̄Π)^{γ_Π} ( (y_t / y_{t-1}) exp(Λ_y) )^{γ_y} ( e_t / e_{t-1} )^{γ_e}.
- Regimes calibrated and described:
  - Flexible Inflation Targeting (FIT): γ_Π determines inflation responsiveness; baseline γ_Π = 1.5000; ̄Π = 1.0122.
  - Strict Inflation Targeting (SIT): γ_Π = ∞ (keeps inflation constant in very short run).
  - Hard Peg (HP): γ_e → ∞ (keeps nominal exchange rate constant); γ_Π = 0.
  - Taylor Rule (TR): responds to inflation and output (baseline γ_y = 0.0000; alternative γ_y = 0.5 considered).
  - Exchange-Rate-Augmented Taylor Rule (ERTR): adds γ_e > 0; example γ_e = 0.5 in alternative experiments.
- Baseline Taylor-rule parameters (selected):
  - ̄Π = 1.0122 (matches average annual inflation of 4.87 percent for disaster-prone countries).
  - ̄Π^∗ = 1.0053 (matches average annual U.S. inflation rate of 2.12 percent).
  - γ_Π = 1.5000; γ_y = 0.0000; γ_R = 0.0000; γ_e = 0.0000.
- Robustness checks: interest rate inertia; replacing CPI inflation with domestic consumption goods inflation (Π^H_t); targeting nominal GDP.

### Disaster shock specification and FIT dynamics (qualitative)
- Disaster affects capital stock and productivity (permanent and stationary components) and shifts export demand via φ_d.
- Key qualitative responses under FIT to an average disaster:
  - Real output and consumption fall initially, then grow faster before converging to balanced growth due to TFP dynamics and rebuilding.
  - Investment increases to rebuild destroyed capital.
  - Exports and imports fall; exports more than imports → net exports contract.
  - Real exchange rate depreciates → curbs import demand, stimulates exports, induces net capital inflow (fall in net foreign assets).
  - Initial nominal exchange rate depreciation raises CPI inflation; domestic goods prices contract and domestic goods inflation remains below steady state for prolonged period, producing CPI inflation undershooting after the initial increase.
  - Central bank targeting CPI inflation adjusts policy rate tracking CPI dynamics; strength of response depends crucially on γ_Π.
  - Welfare-maximizing γ_Π is calibrated near γ_Π ≈ 1.1 in some exercises, implying substantial accommodation of the disaster shock.

---

### Differences across monetary policy regimes — impulse-response and welfare results (baseline calibration)
- Regimes compared: FIT (baseline γ_Π =1.5, γ_y =0, γ_e =0), SIT (γ_Π =∞), HP (γ_Π =0, γ_e →∞), TR (γ_Π =1.5, γ_y =0.5), ERTR (γ_Π =1.5, γ_y =0.5, γ_e =0.5).
- Qualitative regime effects under an average natural disaster:
  - Real effects: disaster shock dominates output; differential impact on output across regimes limited; HP and SIT magnify GDP loss to an extent.
  - Nominal effects: choice of regime materially affects nominal variables (exchange rate, inflation, policy rate).
    - Hard peg: eliminates exchange-rate adjustment → exacerbates recession and causes a fall in CPI inflation; central bank lowers policy rate but initial output loss larger than under FIT.
    - Strict inflation targeting: requires prolonged rate increase to keep inflation constant; exchange rate depreciates less than under FIT.
    - Conventional TR: responsiveness to output leads to accommodation, mildly mitigating output contraction, larger exchange rate depreciation and higher inflation than baseline.
    - ERTR: intermediate responses when central bank reacts to exchange rate.
- Baseline welfare and volatility outcomes (Table 5 — numeric entries preserved):
  - Flexible inflation targeting (γ_Π=1.5, γ_y=0, γ_e=0)
    - Output volatility: 2.8500 (%)
    - Inflation volatility: 1.1403 (%)
    - Welfare level: 0.4611
    - C.E. gain w.r.t. FIT: -
  - Strict inflation targeting (γ_Π=∞, γ_y=0, γ_e=0)
    - Output volatility: 2.8900 (%)
    - Inflation volatility: 0.6173 (%)
    - Welfare level: 0.4596
    - C.E. gain w.r.t. FIT: -0.3253 (%)
  - Hard peg (γ_Π=0, γ_y=0, γ_e=∞)
    - Output volatility: 3.0500 (%)
    - Inflation volatility: 1.2862 (%)
    - Welfare level: 0.4580
    - C.E. gain w.r.t. FIT: -0.6723 (%)
  - Taylor rule (γ_Π=1.5, γ_y=0.5, γ_e=0)
    - Output volatility: 2.9800 (%)
    - Inflation volatility: 1.3083 (%)
    - Welfare level: 0.4575
    - C.E. gain w.r.t. FIT: -0.7807 (%)
  - Exchange-rate-augmented TR (γ_Π=1.5, γ_y=0.5, γ_e=0.5)
    - Output volatility: 2.9900 (%)
    - Inflation volatility: 1.3097 (%)
    - Welfare level: 0.4573
    - C.E. gain w.r.t. FIT: -0.8241 (%)
- Interpretation:
  - Flexible inflation targeting dominates all alternatives in baseline calibration.
  - Strict targeting lowers inflation volatility but slightly raises output volatility and yields welfare loss of -0.3253 percent permanent consumption.
  - Hard peg and Taylor-type rules increase volatilities and yield larger C.E. welfare losses.
  - Positive responsiveness to output (γ_y) or exchange rate (γ_e) reduces welfare relative to FIT.

### Sensitivities and robustness
- Export demand channel (φ_d =0.25 baseline; counterfactual φ_d =0):
  - With φ_d =0 (pure supply shock) results change quantitatively; flexible inflation targeting still dominates.
  - Table 6 (φ_d =0) selected entries (numeric preservation):
    - FIT: Output volatility 2.5100 (%); Inflation volatility 1.1595 (%); Welfare level 0.4566; C.E. gain: -
    - SIT: Output volatility 2.5100 (%); Inflation volatility 0.6264 (%); Welfare level 0.4548; C.E. gain: -0.3942 (%)
    - Hard peg: Output volatility 2.6600 (%); Inflation volatility 1.3057 (%); Welfare level 0.4529; C.E. gain: -0.8103 (%)
    - TR: Output volatility 2.7400 (%); Inflation volatility 1.3426 (%); Welfare level 0.4527; C.E. gain: -0.8541 (%)
    - ERTR: Output volatility 2.7100 (%); Inflation volatility 1.3452 (%); Welfare level 0.4523; C.E. gain: -0.9417 (%)
- Excluding one shock at a time (Subsection 6.2):
  - Flexible inflation targeting remains welfare-maximizing regardless of which shock is deactivated.
  - Selected numeric snapshots (Table 7):
    - Excluding the foreign inflation shock: FIT welfare level 0.4614; SIT 0.4599; HP 0.4584.
    - Excluding the foreign interest rate shock: FIT welfare level 0.4747; SIT 0.4737; HP 0.4723.
    - Excluding the TFP shock: FIT welfare level 0.4918; SIT 0.4909; HP 0.4911.
- CPI inflation targeting versus domestic inflation targeting (Subsection 6.3):
  - Replacing CPI inflation with domestic inflation (Π^H_t) tends to:
    - Stabilize domestic inflation while allowing CPI inflation to rise.
    - Lead the central bank to lower policy rate to avoid fall in domestic inflation; nominal exchange rate increases more; net exports marginally deteriorate; monetary accommodation reduces initial output fall.
  - Table 8 (Domestic inflation targeting) key entries:
    - FIT (domestic targeting): Output volatility 2.6400 (%); Inflation volatility 0.3592 (%); Welfare level 0.4623; C.E. gain: -
    - SIT: Output volatility 2.5800 (%); Inflation volatility 0.0040 (%); Welfare level 0.4619; C.E. gain: -0.0865 (%)
    - Hard peg: Output volatility 3.0500 (%); Inflation volatility 1.2862 (%); Welfare level 0.4580; C.E. gain: -0.9301 (%)
    - TR: Output volatility 2.7700 (%); Inflation volatility 0.4876 (%); Welfare level 0.4595; C.E. gain: -0.6057 (%)
    - ERTR: Output volatility 2.8400 (%); Inflation volatility 0.6675 (%); Welfare level 0.4587; C.E. gain: -0.7787 (%)
  - Conclusion: flexible inflation targeting remains superior even when targeting domestic inflation.

---

### Nominal GDP Targeting (NGT) — implementation and comparative results
- Rationale:
  - NGT targets growth rate of nominal GDP; argued advantages include use of observable variables and determinate equilibrium for any trend inflation.
  - Prior literature suggests NGT can be more effective when supply shocks increase inflation and output variability; small-open-economy exchange rate effects often neglected in prior work.
- Implementation in model:
  - NGT implemented following Garin et al. (2016) as γ_Π → ∞, γ_y → ∞, γ_e = 0.
- Simulation dynamics:
  - NGT is effective at mitigating output collapse after a disaster via accommodating policy, large exchange rate depreciation, and an inflation spike that then returns to steady state (implying a shift in the price level).
  - In the SOE model with demand and supply shocks, NGT entails very large exchange rate and inflation shifts.
- Quantitative comparison (Table 9 entries preserved as presented):
  - Flexible inflation targeting: 1.5002.85001.14030.4611-
  - Nominal GDP targeting: ∞∞03.10001.92320.4486-2.7109
  - Interpretation: NGT increases both output and inflation volatilities relative to FIT and is suboptimal in welfare terms because of large exchange rate shifts and greater inflation volatility.
- Policy implication:
  - Despite supporting output immediately after disasters, NGT’s large exchange rate depreciation and inflation spike raise volatilities and reduce welfare relative to FIT; central banks in small open, disaster-prone economies must account for exchange rate dynamics when evaluating NGT.

---

### Country Q&A excerpts — recurring themes and preserved passages
- Currency arrangements:
  - Mixture of own legal tender, currency unions, and dollarized economies across episodes; pegged regimes and floating regimes documented.
- Monetary policy independence:
  - Independent when own legal tender and no hard peg; not independent for ECCU members or dollarized economies; mixed for soft pegs.
- Typical macro outcomes after disasters (selected verbatim figures and passages preserved):
  - "The economy contracted by 3 percent in 2004..." (Grenada, 2004).
  - "The property damage and the interruption to economic activity are estimated to have totaled 5½ percent of GDP..." (Haiti, 2004).
  - "Real GDP growth declined from 11 percent in 2000 to 5 percent in 2001" and "inflation remained very low at an annual rate of 1¼ percent" (Belize).
  - "the gourde appreciated by about 3½ percent ... and by about 9½ percent in real effective terms..." (Haiti, 1998).
  - "During August–October 2004, interest rates were reduced to 7.6 percent from 13.6 percent..." (Haiti, 2004).
  - "the central bank lowered the policy rate to 6.25 percent..." (Jamaica, 2010).
  - "The ratio of international reserves to money base was 96 percent at end-2015, compared with the statutory mandate of 60 percent..." (Micronesia, 2015).
  - "and raised policy interest rates by 2.5 percentage points..." (Moldova, 2007).
  - "the Central Bank of Sri Lanka (CBSL) raised the policy rate by 25 basis points..." (Sri Lanka, 2016).
- Monetary policy stances and IMF reactions:
  - Tightening, accommodation, and unchanged stances observed depending on exchange rate considerations, reserves, inflation pressures and output losses.
  - IMF advice frequently "Tighten" as recovery firms or to control inflation; sometimes "Accommodate" or "Neutral" when appropriate.
  - Examples of IMF appraisal preserved verbatim in coded excerpts.

*Source: wpiea2022067-print-pdf — Section 2.1 Methodology (narrative analysis design and sample) and related descriptive material from the paper.*

### 2.1   Methodology...................................9

### 2.1   Methodology

### Overview
- The narrative analysis examines macroeconomic and monetary policy performance of countries after climate-related natural disasters, using IMF Article IV staff reports prepared in the year of, and one year following, the occurrence of a disaster.
- The study focuses on disaster-years with annual damages of at least 1 percent of GDP, subject to staff report availability.
- The narrative analysis records: the nature of mobilized monetary policy tools; whether policy was accommodative, tight, or unchanged; IMF staff and/or Board appraisal of the monetary policy stance; and IMF advice for the near future.

### Sample and data
- Final sample: 34 disaster-years, occurring in 16 disaster-prone countries from 1999 to 2017.
- Damage threshold for inclusion: at least 1 percent of GDP in the disaster-year (subject to report availability).
- For countries in currency unions, Article IV staff reports for the union’s central bank were cross-referenced.
- Example country-level facts cited in the analysis:
  - Disaster-prone EMDEs: damages represent approximately 7% of GDP on average versus 0.5% in their peers less exposed to natural disaster.
  - Extreme events causing damages beyond 200% of GDP are documented.
  - Belize: hurricane Keith (October 2000) and hurricane Iris (October 2001) caused damages “of the tune of 30 percent of GDP each,” and GDP growth in 2001 and 2002 was about 8 percentage points lower than in the pre-shock year.

### Narrative analysis design and coding
- Questions in the structured questionnaire (Table 2) include:
  - Does the country have its own legal tender? (Y-N)
  - Is its currency pegged to some other currency or basket of currencies? (Y-N)
  - Can we characterize monetary policy as independent? (Y-N-Mixed)
  - Did GDP contract or slowdown in the aftermath of the disaster? (Y-N-NA)
  - Did inflation increase (or was it expected to increase) in the aftermath of the disaster? (Y-N-NA)
  - Was there an appreciation/depreciation of the real exchange rate? (Appreciation-Depreciation-No change)
  - Were there any challenges for maintaining the peg? (Y-N-NA)
  - Were reserves impacted negatively? (Y-N-NA)
  - Was monetary policy tightened, accommodated or unchanged? (Accommodated-Tightened-Unchanged)
  - What was the monetary policy tool authorities used? (Open question)
  - Did IMF agree with the authorities’ policy action? (Y-N-Neutral)
  - What was the IMF advice on the monetary policy stance to adopt after IMF mission? (Accommodate-Tighten-Neutral)
- Classification of monetary policy independence:
  - “Independent” — full control of monetary policy (own legal tender and policy autonomy).
  - “Not independent” — does not have own legal tender or has a hard peg.
  - “Mixed” — peg or exchange rate anchor but limited capital mobility allows some room for monetary policy.
- Narrative coding procedure involved reading archived Article IV reports and extracting quotes and answers to the questionnaire; Appendix B documents the full narrative analysis with quotes for all disaster-country observations.
- Table 1 provides the list of disasters and corresponding Article IV report years, with cumulative damages (% of GDP) for each observation.

### Link to modeling approach used elsewhere in the paper (methodological context)
- To evaluate optimal monetary policy rules, the paper uses a New-Keynesian model with the Taylor projection solution method.
- Extensions to the baseline model include:
  1. Allowing disaster effects on productivity to have both permanent and temporary components, and allowing disasters to affect export demand.
  2. Introducing a small-economy setup along the lines of Gali and Monacelli (2005).
  3. Considering an array of alternative Taylor-type interest rate rules capturing several monetary policy regimes to evaluate welfare outcomes.

### Key procedural notes and inclusion rules
- Article IV consultations are the source of the staff reports used; due to staff report availability, in a few cases answers are based on consultations occurring two or three years after the disaster (examples cited: El Salvador 2011; Micronesia 2015; Solomon Islands 2014; Samoa 2012).
- For currency-union members (e.g., ECCU), Article IVs for both the country and the union were combined when relevant.
- The narrative analysis restricts the sample to cases that suffered a cumulative loss of at least 5 percent of GDP for certain distributional figures (e.g., Figure 2), where noted in the text.

*Source: wpiea2022067-print-pdf — Section 2.1 Methodology (narrative analysis design and sample) and related descriptive material from the paper.*

### 2.2   Results

### 2.2   Results

### Main empirical findings on macroeconomic impact
- The narrative analysis covers weather-related natural disasters with associated damages of at least 1% of GDP, leading to a sample of 34 incidents that occurred in 16 countries.
- In most cases, GDP growth declined in the aftermath of disasters.
- Inflation often increased, although there are counterexamples (e.g., Belize): "Real GDP growth declined from 11 percent in 2000 to 5 percent in 2001" and "inflation remained very low at an annual rate of 1¼ percent".
- Most disasters occurred in countries with at least partial control of their monetary policy; only 12 percent of cases refer to disasters occurred in countries with dollarized economies.
- If non-pegged countries were considered, the percentage of countries that experienced an impact on their reserves would go down to 35 percent.

### Monetary policy stance adopted after disasters
- The monetary policy stance was changed in virtually all cases where there was room for maneuver, highlighting the perceived importance of monetary policy as a tool for mitigating adverse effects of natural disasters.
- When changed, the monetary policy stance was:
  - Tightened in slightly more than half of the cases (almost 56 percent of disasters).
  - Accommodated in the remaining cases.
- The choice between tightening and accommodation signals heterogeneous weights given to inflation and exchange rate considerations versus output losses.
- The main monetary policy tool utilized in the aftermath of disasters was the interest rate, although other tools (such as the money supply) were mobilized in several cases.
- Constraints to changes in the monetary policy stance are typically hard pegs or dollarized economies.

### IMF staff appraisal and advice
- IMF staff and/or directors generally agreed with authorities when they adopted a tight monetary policy stance.
- Directors also agreed with loosening in a number of cases: "about half of instances in which authorities adopted a loose monetary policy stance".
- IMF staff appraised positively some accommodative monetary responses, indicating that in certain cases expert judgment emphasized stimulating the economy to avoid recession traps.
- There are no cases where the IMF advice was to switch from a tight to a loose monetary policy stance in the near future, while the reverse (advice to tighten after a loose stance) does occur. This outcome is likely driven by concerns about:
  - inflation derailment and anchoring of expectations,
  - central bank credibility,
  - availability of reserves,
  - exchange rate stabilization.

### Robustness notes and limitations from staff consultations
- The narrative results were discussed with IMF mission chiefs and desk economists covering about a third of the countries in the sample; teams found the findings generally in line with field experience and emphasized monetary policy dilemmas posed by natural disasters.
- Teams noted omissions in the study:
  - exclusion of the fiscal response or donor support,
  - exclusion of lower-impact disasters (damages smaller than 1 percent of GDP), which are high-frequency and largely unreported in IMF staff reports and therefore excluded for feasibility reasons.

### Transition to analytical framework
- The heterogeneity in monetary policy conduct and advice, along with the multiple stabilization challenges triggered by natural disasters (inflation, business cycle, exchange rate), motivate the quantitative investigation using the model presented in the next section: a SOE New-Keynesian model with stochastic trend growth and disaster shocks, Epstein-Zin preferences, disaster effects on capital and TFP, and alternative Taylor-type interest rate rules.

*Sources: IMF staff reports and authors’ calculations.*

### 3.4   Monetary Policy

### 3.4   Monetary Policy

### Monetary policy feedback rule and regimes
- Central bank sets the interest rate according to the generalized feedback rule:
  R_t / R = (Π_t / ̄Π)^{γ_Π} ( (y_t / y_{t-1}) exp(Λ_y) )^{γ_y} ( e_t / e_{t-1} )^{γ_e}.
- Alternative calibrated monetary policy regimes explored (each obtained by appropriate parametrization):
  - Flexible Inflation Targeting (FIT)
    - Central bank concerned exclusively with inflation stabilization, allowing temporary deviations from the inflation objective (Svensson, 2000).
    - The responsiveness parameter γ_Π determines how quickly inflation returns to target; γ_Π close but above 1 satisfies the Taylor principle while keeping policy mild.
    - Rule (limiting formulation): R_t / R = (Π_t / ̄Π)^{γ_Π}.
  - Strict Inflation Targeting (SIT)
    - Limiting case with γ_Π = ∞; central bank keeps inflation constant (stabilized in the very short run).
    - Rule: R_t / R = (Π_t / ̄Π)^{γ_Π}, γ_Π = ∞.
  - Hard Peg (HP)
    - Central bank keeps nominal exchange rate constant by setting γ_e → ∞.
    - Rule: R_t / R = ( e_t / e_{t-1} )^{γ_e}.
  - Taylor Rule (TR)
    - Responds to both inflation and economic activity; includes responsiveness γ_y to output growth (Fernández-Villaverde and Levintal, 2018).
    - Rule: R_t / R = (Π_t / ̄Π)^{γ_Π} ( (y_t / y_{t-1}) exp(Λ_y) )^{γ_y}.
  - Exchange-rate-augmented Taylor Rule (ERTR)
    - TR extended to respond to nominal exchange rate changes (γ_e > 0), capturing welfare concerns from depreciations increasing import prices.
    - Rule: R_t / R = (Π_t / ̄Π)^{γ_Π} ( (y_t / y_{t-1}) exp(Λ_y) )^{γ_y} ( e_t / e_{t-1} )^{γ_e}.
- Robustness checks (Section 6) include:
  - Adding interest rate inertia.
  - Replacing CPI inflation responsiveness (Π_t) with inflation of domestic consumption goods prices (Π^H_t).
  - Targeting nominal GDP.

### Baseline monetary policy calibration (selected parameters and context)
- Inflation target parameter: ̄Π = 1.0122 (matches average annual inflation of 4.87 percent for disaster-prone countries).
- Steady state of foreign inflation: ̄Π^∗ = 1.0053 (matches average annual U.S. inflation rate of 2.12 percent).
- Inflation parameter in Taylor rule: γ_Π = 1.5000 (satisfies the Taylor principle).
- Output growth parameter in Taylor rule: γ_y = 0.0000 (baseline shuts down output responsiveness).
- Interest rate smoothing in Taylor rule: γ_R = 0.0000.
- Exchange rate parameter in Taylor rule: γ_e = 0.0000.
- Note: remaining Taylor-rule parameters are set to zero in the baseline illustrative results; alternative policy experiments activate these objectives.

### Interaction with disaster shocks and small-open economy dynamics (under FIT)
- Disaster shock specification and impacts:
  - Disaster affects stock of capital and productivity (both permanent and stationary components).
  - Export demand is downward-shifted by disasters through parameter φ_d.
- General equilibrium and external sector:
  - Imports: imp_t = c^F_t + x^F_t = (1−φ) (s_t)^{−χ_c} (c_t + x_t).
  - Exports: exp_t = φ^∗ ( (p^H_t e_t) / p^∗_t )^{−χ^∗_c} y^∗_t − ψ_d d_t θ_t (ψ_d governs disaster impact on external demand).
  - Balance of payments (real terms): ̃p^H_t exp_t − s_t imp_t + s_t ( R^∗_{t−1} Ψ_{t−1}^{−1} ̃b^∗_{t−1} / Π^∗_t ) = s_t ( ̃b^∗_t − ̃b^∗_{t−1} / Π^∗_t ).
- Impulse response setup:
  - Stochastic steady state obtained by simulating model without shocks for 100 quarters.
  - A one-off average disaster shock is applied and IRFs traced for 20 quarters.
  - Output, consumption, investment, exports, imports, net exports and net foreign assets are non-stationary and plotted in percent deviations from the stochastic steady state (they grow each period at the same growth rate as TFP).
- Key qualitative dynamics following an average disaster under FIT:
  - Real output and consumption fall initially, then grow faster before converging to balanced growth path (due to disaster hitting both TFP components and subsequent rebound).
  - Investment increases to rebuild destroyed capital stock.
  - Exports and imports fall; the fall in exports is more pronounced than imports, causing net exports to contract.
  - Real exchange rate depreciates (increase) to curb import demand, stimulate exports and induce net capital inflow (fall in net foreign assets).
  - Initial sharp nominal exchange rate depreciation raises CPI inflation; domestic goods prices contract and domestic goods inflation remains below steady state for prolonged period, producing CPI inflation undershooting after the initial increase.
  - Central bank targeting CPI inflation adjusts policy rate tracking CPI dynamics; strength of response depends crucially on γ_Π.
    - Example calibration note: setting γ_Π = 1.1 (welfare-maximizing response subject to Taylor principle) implies substantial accommodation of the disaster shock; CPI inflation and policy rate responses resemble a Taylor rule that responds to output and GDP fall is mitigated.

*Source: 3.4 Monetary Policy, wpiea2022067-print-pdf*

### 5.2   Differences Associated with Alternative Monetary Policy Regimes

### 5.2   Differences Associated with Alternative Monetary Policy Regimes

### Regimes compared and impulse-response patterns
- Monetary policy regimes analyzed: flexible inflation targeting (baseline, γ_Π =1.5, γ_y =0, γ_e =0), strict inflation targeting (γ_Π →∞, γ_y =0, γ_e =0), hard peg (γ_Π =0, γ_y =0, γ_e →∞), conventional Taylor rule (γ_Π =1.5, γ_y =0.5, γ_e =0), and exchange-rate-augmented Taylor rule (γ_Π =1.5, γ_y =0.5, γ_e =0.5).
- Under an average natural disaster shock:
  - Real effects: Differential impact on output across regimes is limited because the natural disaster shock dominates output dynamics; both a hard peg and strict inflation targeting magnify the GDP loss to an extent.
  - Nominal effects: The choice of regime materially affects nominal variables.
    - Hard peg: eliminates the shock-absorbing effect of the exchange rate, exacerbates the recession and causes a fall in aggregate demand and CPI inflation; the central bank lowers the policy rate but initial output loss is larger than under flexible inflation targeting.
    - Strict inflation targeting: requires a more prolonged increase in the interest rate to keep inflation constant; exchange rate still depreciates but less than under flexible inflation targeting.
    - Conventional Taylor rule (reaction to inflation and output): responsiveness to output leads to monetary policy accommodation, mildly mitigating the output contraction, larger exchange rate depreciation and higher inflation than baseline.
    - Exchange-rate-augmented Taylor rule: intermediate responses between flexible inflation targeting and conventional Taylor rule when the central bank also reacts to the exchange rate.

### Key numerical outcomes from baseline welfare simulations (Table 5)
- Notes: Output and inflation volatilities are standard deviations of percent fluctuations around trends, simulated for 900 quarters after 100 quarters with no shocks. Welfare level is average of simulated recursive households’ utility. C.E. gain is permanent percent consumption change needed to match FIT (minus sign = welfare loss).
- Flexible inflation targeting (γ_Π=1.5, γ_y=0, γ_e=0)
  - Output volatility: 2.8500 (%)
  - Inflation volatility: 1.1403 (%)
  - Welfare level: 0.4611
  - C.E. gain w.r.t. FIT: -
- Strict inflation targeting (γ_Π=∞, γ_y=0, γ_e=0)
  - Output volatility: 2.8900 (%)
  - Inflation volatility: 0.6173 (%)
  - Welfare level: 0.4596
  - C.E. gain w.r.t. FIT: -0.3253 (%)
- Hard peg (γ_Π=0, γ_y=0, γ_e=∞)
  - Output volatility: 3.0500 (%)
  - Inflation volatility: 1.2862 (%)
  - Welfare level: 0.4580
  - C.E. gain w.r.t. FIT: -0.6723 (%)
- Taylor rule (γ_Π=1.5, γ_y=0.5, γ_e=0)
  - Output volatility: 2.9800 (%)
  - Inflation volatility: 1.3083 (%)
  - Welfare level: 0.4575
  - C.E. gain w.r.t. FIT: -0.7807 (%)
- Exchange-rate-augmented TR (γ_Π=1.5, γ_y=0.5, γ_e=0.5)
  - Output volatility: 2.9900 (%)
  - Inflation volatility: 1.3097 (%)
  - Welfare level: 0.4573
  - C.E. gain w.r.t. FIT: -0.8241 (%)

### Welfare implications and interpretation
- Flexible inflation targeting dominates all other regimes under the baseline calibration.
- Strict inflation targeting: lowers inflation volatility but slightly raises output volatility and yields a welfare loss equivalent to about -0.3253 percent permanent consumption.
- Hard peg: removes exchange-rate shock absorption, increases output and inflation volatilities, and yields a C.E. welfare loss of -0.6723 percent.
- Conventional and exchange-rate-augmented Taylor rules: increase output and inflation volatilities and yield C.E. welfare losses of about -0.7807 percent and -0.8241 percent, respectively.
- Welfare-maximizing responsiveness to inflation under baseline is small (γ_Π ≈1.1), implying it is optimal for the central bank to minimally respond to inflation to keep expectations anchored while largely accommodating disaster shocks.
- Positive monetary policy responsiveness to output (γ_y) or the exchange rate (γ_e) reduces welfare relative to flexible inflation targeting.

### Sensitivity: export demand channel (Subsection 6.1) and results without export channel
- Baseline includes export demand channel (φ_d =0.25) to capture tourism-dependent dynamics; counterfactual sets φ_d =0.
- With φ_d =0 (disaster = pure supply shock):
  - Domestic inflation initially falls due to nominal exchange rate appreciation (decrease) and later overshoots because domestic goods inflation stays elevated; policy rate tracks CPI inflation.
  - Export channel amplifies real effects but does not change sign of responses; it is important for nominal variable responses, especially CPI inflation.
- Table 6 (No export demand channel, φ_d =0) key entries:
  - Flexible inflation targeting (γ_Π=1.5, γ_y=0, γ_e=0): Output volatility 2.5100 (%); Inflation volatility 1.1595 (%); Welfare level 0.4566; C.E. gain: -
  - Strict inflation targeting (γ_Π=∞): Output volatility 2.5100 (%); Inflation volatility 0.6264 (%); Welfare level 0.4548; C.E. gain: -0.3942 (%)
  - Hard peg: Output volatility 2.6600 (%); Inflation volatility 1.3057 (%); Welfare level 0.4529; C.E. gain: -0.8103 (%)
  - Taylor rule: Output volatility 2.7400 (%); Inflation volatility 1.3426 (%); Welfare level 0.4527; C.E. gain: -0.8541 (%)
  - Exchange-rate-aug. TR: Output volatility 2.7100 (%); Inflation volatility 1.3452 (%); Welfare level 0.4523; C.E. gain: -0.9417 (%)
- Welfare ranking remains unaltered: flexible inflation targeting continues to dominate.

### Sensitivity: excluding one shock at a time (Subsection 6.2)
- Exercise: deactivate one shock at a time (while keeping natural disasters and other shocks active) and compute welfare across regimes.
- Finding: Regardless of which shock is deactivated (examples reported include excluding foreign inflation shock, foreign interest rate shock, foreign demand shock, or TFP shock), flexible inflation targeting remains the welfare-maximizing regime.
- Selected numeric outcomes (from Table 7 snapshots):
  - Excluding the foreign inflation shock: Flexible inflation targeting welfare level 0.4614; strict targeting welfare level 0.4599; hard peg welfare level 0.4584.
  - Excluding the foreign interest rate shock: Flexible inflation targeting welfare level 0.4747; strict targeting welfare level 0.4737; hard peg welfare level 0.4723.
  - Excluding the TFP shock: Flexible inflation targeting welfare level 0.4918; strict targeting welfare level 0.4909; hard peg welfare level 0.4911.
- Conclusion: Results do not hinge on any single shock.

### CPI inflation targeting versus domestic inflation targeting (Subsection 6.3)
- Alternative rule: replace CPI inflation (Π_t) with domestic inflation (Π^H_t) as the target variable.
- Impulse-response differences:
  - Targeting domestic inflation stabilizes domestic inflation while allowing CPI inflation to rise.
  - Central bank lowers policy rate to avoid fall in domestic inflation; nominal exchange rate increases more than under CPI targeting; real exchange rate appreciates only slightly more; net exports marginally deteriorate; monetary accommodation reduces initial output fall.
- Table 8 (Domestic inflation targeting) key entries:
  - Flexible inflation targeting (γ_Π=1.5, γ_y=0, γ_e=0): Output volatility 2.6400 (%); Inflation volatility 0.3592 (%); Welfare level 0.4623; C.E. gain: -
  - Strict inflation targeting (γ_Π=∞): Output volatility 2.5800 (%); Inflation volatility 0.0040 (%); Welfare level 0.4619; C.E. gain: -0.0865 (%)
  - Hard peg: Output volatility 3.0500 (%); Inflation volatility 1.2862 (%); Welfare level 0.4580; C.E. gain: -0.9301 (%)
  - Taylor rule: Output volatility 2.7700 (%); Inflation volatility 0.4876 (%); Welfare level 0.4595; C.E. gain: -0.6057 (%)
  - Exchange-rate-aug. TR: Output volatility 2.8400 (%); Inflation volatility 0.6675 (%); Welfare level 0.4587; C.E. gain: -0.7787 (%)
- Targeting domestic inflation generally reduces output and inflation volatilities and raises welfare relative to CPI inflation targeting; welfare losses relative to FIT are smaller (except for hard peg).
- Crucial result preserved: flexible inflation targeting remains superior to alternative regimes even when targeting domestic inflation.

_Italic: Content based on "5.2   Differences Associated with Alternative Monetary Policy Regimes" and related subsections from the provided chapter._

### 6.4   Nominal GDP Targeting

### 6.4   Nominal GDP Targeting

### Objective and theoretical appeal
- Nominal GDP targeting (NGT) targets the growth rate of nominal GDP and has been argued to offer advantages relative to inflation targeting:
  - Requires knowledge of easily observable variables instead of, e.g., the output gap.
  - Does not suffer from indeterminacy issues because, in the long run, NGT is equivalent to price level targeting, supporting a determinate equilibrium for any level of trend inflation.
  - McKibbin et al. (2021) argue that NGT can be more effective when climate change increases the variability of inflation and output due to more frequent supply shocks.
- Prior contributions generally neglect effects of NGT on exchange rate dynamics; results may not extend to a small-open-economy setting.
- Jensen (2002) and Billi (2017) show the desirability of NGT arises only in the presence of supply shocks, i.e., when the central bank faces a trade-off between stabilizing inflation and output.

### Implementation in the model
- The paper follows Garin et al. (2016) in choosing a parametrization of the Taylor rule to obtain NGT:
  - The rule specification is presented as: R_t / R =  Π_t ̄Π  (γ_Π)  [ y_t / y_{t−1} exp(Λ_y ) ]^{γ_y}, γ_Π,γ_y = ∞,,γ_y = ∞.(44)
  - Nominal GDP targeting is implemented with γ_Π → ∞, γ_y → ∞, γ_e = 0.

### Simulation dynamics and mechanisms
- By targeting nominal GDP growth, NGT:
  - Is very effective at mitigating the output collapse in the aftermath of the disaster realization.
  - Achieves this outcome through an accommodating monetary policy, a large exchange rate depreciation, and a spike in inflation that then returns to steady state—essentially implying a shift in the price level.
- However, in the small-open-economy model with both demand and supply shocks, NGT entails very large shifts in the exchange rate and thus in inflation.

### Quantitative comparisons (as reported in Table 9)
- The paper compares the baseline flexible inflation targeting (FIT) regime with NGT. The reported model figures are presented exactly as in the source:
  - Flexible inflation targeting: 1.5002.85001.14030.4611-
  - Nominal GDP targeting: ∞∞03.10001.92320.4486-2.7109
- Interpretation given in the text:
  - NGT substantially increases both output and inflation volatilities relative to FIT.
  - As a result, NGT is suboptimal relative to FIT in welfare terms; one reason is that NGT entails too large exchange rate shifts and hence greater inflation volatility.

### Policy-relevant findings and implications
- Despite NGT’s effectiveness at supporting output immediately after a disaster, the accompanying large exchange rate depreciation and inflation spike raise output and inflation volatilities and reduce welfare relative to flexible inflation targeting.
- In the model with disaster shocks, flexible inflation targeting—allowing inflation to depart temporarily from target—is superior to NGT and other alternatives because it keeps the interest rate response at the lowest level needed to keep inflation expectations anchored while minimizing the impact on real activity.
- The results underscore that central banks in small open, disaster-prone economies must account for exchange rate dynamics when evaluating NGT; avoiding direct targeting of the exchange rate or economic activity remains welfare-improving.

*Source: 6.4 Nominal GDP Targeting, wpiea2022067-print-pdf*

### References

### References

### Bibliographic references and research themes
- Collection of cited works covering:
  - Tropical cyclones and post-disaster reconstruction (Adam and Bevan (2020)).
  - Current account norms in resource-rich economies (Araujo et al. (2016)).
  - Rare disasters, asset markets, and welfare costs (Barro (2006), Barro (2009), Barro (2015)).
  - Monetary policy rules for open economies and exchange rate dynamics (Batini et al. (2003); Benigno (2004); Gali and Monacelli (2005); Svensson (2000); Kollmann (2002)).
  - Nominal GDP targeting and interest-rate rules (Billi (2017); Bullard and Singh (2020); Garin, Lester, and Sims (2016); Giannoni (2014); Jensen (2002); McCallum and Nelson (1999); Taylor (1993)).
  - Disaster risk, macroeconomic outcomes, and climate-related shocks (Cantelmo (2022); Cantelmo, Melina, and Papageorgiou (2019); Burke et al. (2015); Hsiang and Jina (2014); Gourio (2012); De Winne and Peersman (2021); McKibbin et al. (2021); Nordhaus (2019); Intergovernmental Panel on Climate Change (2018)).
  - Solution and computation methods for DSGE models with rare disasters and stochastic volatility (Caldara et al. (2012); Fernández-Villaverde and Levintal (2018); Levintal (2018); Judd (1992)).
  - Empirical studies on natural disasters' effects on behavior, risk preferences, tourism, and growth (Cameron and Shah (2015); Cassar, Healy, and von Kessler (2017); Brown et al. (2018); Rossello, Becken, and Santana-Gallego (2020); Okafor, Adeola, and Folarin (2021); Fiala (2017); van den Berg, Fort, and Burger (2009)).

### Appendix A — List of Disaster-Prone Countries (Table A.1)
- Title: Disaster-Prone Countries: Fourth Quartile (75%-100%) of the Annual Probability Distribution of Natural Disasters.
- Table columns: Country | Annual Probability per 1000 sq. km | Damages (% of GDP) Average | Damages (% of GDP) Max | Small economy.
- Key table characteristics and metadata:
  - Countries are ordered by the annual probability of a natural disaster per 1000 squared kilometers over the sample 1998-2017.
  - EM-DAT provides damages in US dollars. Damages in percent of GDP are obtained dividing damages by GDP of the year of the event.
  - Damages (% of GDP) are computed for each country by using data for each single event over the sample 1998-2017.
  - Small economies comprise small states and low-income countries.
  - Footnotes:
    - ∗ Denotes Small states which are countries with a population below 1.5 million that are not advanced economies or high-income oil exporting countries (IMF).
    - ∗∗ Denotes Low-income-countries which are countries with a GNI per capita below $995 in 2017 (World Bank).
- Sample rows (as presented in the table):
  - Marshall Islands | 100.00 | 2.72 | 2.72 | Yes ∗
  - St. Vincent and the Grenadines | 100.00 | 4.57 | 15.0 | Yes ∗
  - Tuvalu | 100.00 | N.A. | N.A. | Yes ∗
  - Micronesia, Fed. Sts. | 50.00 | 1.85 | 3.49 | Yes ∗
  - St. Lucia | 48.39 | 1.07 | 3.13 | Yes ∗
  - Tonga | 46.67 | 12.22 | 29.0 | Yes ∗
  - Grenada | 44.12 | 74.8 | 148 | Yes ∗
  - Dominica | 33.33 | 1118 | 260 | Yes ∗
  - Kiribati | 24.69 | N.A. | N.A. | Yes ∗
  - Maldives | 16.67 | N.A. | N.A. | Yes ∗
  - (Additional entries continue in the table through Moldova | 1.33 | 2.47 | 9.22 | No)
- Source notation in table: Source: Cantelmo et al. (2019).

### Appendix B — Narrative Analysis Documentation (Table B.1)
- Table B.1 structure: For each country-year episode, 12 standardized questions with recorded answers and direct quotes from the corresponding Article IV Staff Report.
- Question set includes:
  1. Does the country have its own legal tender?
  2. Is its currency pegged to some other currency or basket of currencies?
  3. Can we characterize monetary policy as independent?
  4. Did GDP contract or slowdown in the aftermath of the disaster?
  5. Did inflation increase (or was it expected to increase) in the aftermath of the disaster?
  6. Was there an appreciation/depreciation of the real exchange rate?
  7. Were there any challenges for maintaining the peg? (peg countries)
  8. Were reserves impacted negatively?
  9. Was monetary policy tightened, accommodated or unchanged?
  10. What was the monetary policy tool authorities used?
  11. Did IMF agree with the authorities’ policy action?
  12. What was the IMF advice on the monetary policy stance to adopt after IMF mission?
- Selected country-year highlights with verbatim numeric and textual reporting from the staff reports:
  - Belize, 2000 (quotes from 2001 Article IV Staff Report)
    - Legal tender: Y [Belize dollar]
    - Peg: Y ...Belize dollar, pegged to the U.S. dollar since 1976...
    - Monetary policy independence: Mixed [Belize has a soft peg with some room for independent monetary policy]
    - GDP: N ...Real GDP growth is estimated to have increased sharply to 10.4 percent in 2000...
    - Inflation: Y ...Consumer prices increased by 1 percent during the year...
    - REER: Appreciation ...The Belize dollar appreciated in real effective terms by about 4 percent from end-1998 to end-2000...
    - Reserves: Y ...and a similar loss of net international reserves as the central bank absorbs liquidity...
    - Monetary policy stance: Tightened ...reduce excess liquidity to help secure the sustainability of the exchange rate peg...
    - IMF view: Y ...The staff welcomes the authorities’ decision...
  - Belize, 2001 (quotes from 2002 Article IV Staff Report)
    - GDP: Y ...Real GDP growth declined from 11 percent in 2000 to 5 percent in 2001...
    - Inflation: N ...inflation remained very low at an annual rate of 1¼ percent...
    - REER: Appreciation ...appreciated by about 5 percent...
    - Monetary policy: Accommodated ...a sizeable liquidity overhang accumulated...
    - IMF advice: Tighten ...The staff also recommended a tightening of monetary policy...
  - Belize, 2007 (quotes from 2008 Article IV Staff Report)
    - GDP: Y ...economic growth weakened in 2007...
    - Inflation: N ...inflation remained low at 3 percent...
    - Reserves: N ...international reserves increased further, to US$108 million by end-2007...
    - Monetary policy: Unchanged
  - Dominica, 2015 (quotes from 2016 Article IV Staff Report)
    - Legal tender: N [Eastern Caribbean dollar]
    - Monetary policy independence: N ...the monetary policy stance is decided by the Monetary Council of the ECCU...
    - GDP: Y ...Economic growth contracted by nearly 4 percent last year...
    - REER: Appreciation ...REER appreciated by 4.8 percent...
    - Reserves: N ...ratio of international reserves to money base was 96 percent at end-2015...
    - Monetary policy: NA [Monetary policy is not independent]
  - Dominica, 2017 (quotes from 2018 Article IV Staff Report)
    - GDP: Y ...While Erika had caused severe damage, estimated at 96 percent of GDP...
    - Inflation: Y ...Sharp increases in the prices of food and medication...
    - REER: Appreciation ...REER depreciated in 2017...
    - Reserves: Y ...reserves decreased modestly in 2017...
  - Dominican Republic, 1998 (quotes from 1999 Article IV Staff Report)
    - Legal tender: Y [Dominican Peso]
    - Exchange rate regime: N ...flexible exchange rate policy...
    - Monetary policy independence: Y ...Monetary policy is conducted through a mix of direct and indirect instruments...
    - GDP: Y ...real GDP growth slowed modestly...
    - Inflation: Y ...inflation rose...
    - REER: Depreciation ...5% percent depreciation of the real effective exchange rate during 1998...
    - Reserves: N ...International reserves rose by about US 100 million...
    - Monetary policy: Accommodated ...Base money growth (12-month basis) accelerated to nearly 20 percent...
    - IMF advice: Tighten ...Policy discussions focused on the need to ... through a tighter stance and a prudent monetary policy...
  - El Salvador, 1998 (quotes from 1999 Article IV Staff Report)
    - Legal tender: Y [El Salvador Colón]
    - Peg: Y ...the present peg to the U.S. dollar will be maintained...
    - GDP: Y ...slowdown in real GDP growth to 2.2% percent...
    - Inflation: Y ...consumer prices rose by 4.2 percent (3% percent in the program) due to the impact of Hurricane Mitch...
    - REER: Appreciation ...The colon appreciated by 2.7 percent in real effective terms...
    - Monetary policy: Tightened ...The rate of growth of broad money declined to 9% percent in 1998...
  - El Salvador, 2005 / 2009 / 2011 (quotes from 2006, 2010, 2012 Article IV Staff Reports)
    - Legal tender status across episodes: N [U.S. Dollar] — dollarization / no separate legal tender category.
    - Monetary policy independence: N ...lack of independent monetary policy...
    - GDP and inflation outcomes vary by episode (examples: 2009 GDP fell 0.5 percent (y/y) in Q1 2010 after declining 4.9 percent in the last quarter of 2009; 2011 GDP noted as low growth).
    - Reserves: mixed (2009: Y ...decrease in net international reserves ...)
  - Fiji, multiple episodes (quotes from 2004, 2010, 2011, 2013, 2017 Article IV Staff Reports)
    - Legal tender: Y [Fijian Dollar]
    - Peg/anchor: Y ...linked to a basket of currencies of Fiji’s five major trading partners...
    - Monetary policy independence: Mixed ...fiscal and monetary policy must be consistent with the peg...
    - GDP outcomes: episodes include no contraction (2003/2004: N ...growth high by historical standards) and contractions (2010: Y ...contracted by 3 percent; 2009: Y ...expected to contract by 2½ percent in 2009).
    - Inflation: examples include N ...inflation remained modest; Y ...inflation was on a declining trend in 2012 reaching 2.5 percent...
    - Reserves: mixed (2004: Y ...International reserves have declined relative to imports; 2011: N ...Foreign exchange reserves have improved steadily...)
    - Monetary policy actions: examples include Tightened (2004: Increased interest rates ...RBF raised interest rates by 50 basis points), Unchanged (2009), Accommodated with decreased policy rates (2011), and guidance to Tighten in staff recommendations.
- Table B.1 contains extensive country-episode level coding and verbatim quotes for many additional countries and years beyond the samples above.

*Italic: Source: wpiea2022067-print-pdf - References (https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022067-print-pdf.pdf).*

### 3.8 percent in 2017 from 0.4 percent in 2016...

### Monetary Policy in Disaster-Prone Developing Countries — selected country Q&A excerpts

### Overview: recurring themes across country case notes
- Currency arrangements:
  - Countries either have their own legal tender (e.g., Haitian Gourde, Jamaican Dollar, Mauritian Rupee, Moldovan Leu, Samoan Tala, Solomon Island Dollar, Sri Lankan Rupee) or use/are members of currency unions or foreign legal tender (e.g., U.S. Dollar in Marshall Islands and FSM; Eastern Caribbean dollar in ECCU members).
  - Pegging: repeated references to pegged regimes (Eastern Caribbean dollar pegged to the U.S. dollar; Samoan tala pegged to a basket) and to fully floating regimes (e.g., Moldova).
- Monetary policy independence:
  - Characterized as independent where the country has its own legal tender and no peg (e.g., Haiti, Jamaica, Moldova).
  - Not independent for ECCU members or dollarized economies (e.g., Grenada, St. Vincent and the Grenadines, Marshall Islands, Micronesia).
  - Mixed cases where soft pegs allow some room for policy (e.g., Samoa, Solomon Islands).
- Common IMF assessments and advice:
  - IMF often agreed with accommodative or tightening actions depending on circumstances.
  - Typical IMF advice after missions: "Tighten" as recovery firms or to control inflation; sometimes "Accommodate" or "Neutral" when appropriate.

### Macroeconomic impacts after disasters (GDP, inflation, real exchange rate)
- GDP outcomes:
  - Y[GDP contracted or slowed] noted for many countries after disasters: Grenada (2004) — "The economy contracted by 3 percent in 2004..."; Haiti (2004) — "The property damage and the interruption to economic activity are estimated to have totaled 5½ percent of GDP..."; Jamaica (2004) — "Real GDP contracted sharply in late 2004..."; Marshall Islands (2015) — "overcoming the contraction of the previous year..."; Solomon Islands (2014) — "The impact of Cyclone Raquel and El Niño has caused a reduction in agricultural production...".
  - N[No contraction or slowdown] in other cases: Grenada (1999) — "reflecting the rapid growth in activity, the average per capita income rose..."; Haiti (1998) — "Haitian authorities have managed to maintain macroeconomic stability..."; Sri Lanka (2016) — "Growth has held up despite severe weather...".
- Inflation outcomes:
  - Y[Inflation increased] examples: Fiji (text excerpt: "3.8 percent in 2017 from 0.4 percent in 2016..." indicating a pickup); Jamaica (2004) — "Consumer prices registered a marked increase..."; Haiti (2012) — "Inflation spiked but remained at single digits..."; Sri Lanka (2016) — "inflation has picked up..."; Moldova (2007) — "inflation continues to be stubbornly in double digits...".
  - N[Inflation did not increase or remained low]: Grenada (1999, 2004) — "low inflation", Mauritius (2002) — "Consumer price inflation has recently shown a declining trend...", Samoa (2012) — "inflation remains subdued...".
  - Volatility examples: Haiti (2004) — "prices have been highly volatile on a month-to-month basis, as a result of supply disruptions..."
- Real exchange rate (REER/NEER) movements:
  - Appreciation: multiple instances — Fiji 2016: "NEER and REER appreciated..." ; Grenada (1999) — "The real effective value of the Eastern Caribbean dollar ... has appreciated..."; Haiti (1998) — "the gourde appreciated by about 3½ percent ... and by about 9½ percent in real effective terms..."; Micronesia (2015) — "the real effective rate appreciated by 5 percent...".
  - Depreciation: Grenada (2004) — "real effective exchange rate (REER) has depreciated ..."; Haiti (2004) — "both nominal and real exchange rates have depreciated..."; Jamaica (2010) — "to a depreciation of the currency in real effective terms..."; St. Vincent and the Grenadines (2010, 2013) — "Real effective exchange rate depreciation...".
  - No change: Samoa (2012) — "The exchange rate has remained stable..."; Solomon Islands (2014) — "the exchange rate has been stable..."; Moldova (2007) — "the exchange rate remains broadly in line with fundamentals...".

### Reserves and peg maintenance
- Peg maintenance challenges:
  - NA or no reported challenges for many pegged regimes: ECCU members — "unchanged peg to the U.S. dollar since 1976..."; Samoa/Solomon Islands — "pegged ... has remained broadly stable..."
  - Reported reserve pressures: Sri Lanka (2016) — "international reserves hit their bottom..."; Jamaica (2007) — "to stem reserve losses..."; Solomon Islands (2014) — "FX reserves could diminish in the future..."
- Reserve outcomes:
  - Reserves increased or remained adequate in several cases: Grenada (2004) — "Gross international reserves ... have continued to rise..."; Haiti (1998, 2004, 2012) — "official reserves have risen..."; Micronesia (2015) — "ratio of international reserves to money base was 96 percent at end-2015, compared with the statutory mandate of 60 percent..."; Samoa (2012) — "reserves are adequate...".
  - Reserve declines noted: Jamaica (2010) — "Net international reserves fell..."; Fiji (implied but text says "foreign reserves remained adequate..." despite negative impact).

### Monetary policy stance, instruments used, and IMF reaction
- Stance categorizations and examples:
  - Tightened:
    - Moldova (2000): "Monetary policy was successfully tightened in the first half of 2000..."
    - Moldova (2007): "Despite the tightening ... the central bank ... raised policy interest rates by 2.5 percentage points..."
    - Jamaica (2004, 2007): "Monetary policy has been geared at containing inflation..." / "Tighten monetary policy moderately..."
    - Sri Lanka (2016): "the Central Bank of Sri Lanka (CBSL) raised the policy rate by 25 basis points..."
    - IMF advice often: "Tighten" — "Monetary policy should be tightened as the recovery becomes firmer..." ; "Monetary policy should be tightened further..."
  - Accommodated:
    - Fiji (post-disaster context): "Monetary policy remains accommodative..." and authorities "Decreased interest rates" — "The combination of lower lending interest rates and..."
    - Samoa (2012): "A loose monetary policy has supported the recovery..." and instrument "Decrease interest rates" — "lower interest rates..."
    - Solomon Islands (2014): "Monetary policy remains accommodative..."
    - IMF advice sometimes: "Accommodate" or to "stand ready to adopt a tightening stance..."
  - Unchanged / Neutral:
    - Haiti (2012): "Directors endorsed the current neutral stance of monetary policy..." and IMF advice: "Tighten... encouraged the authorities to keep price inflation in check..."
    - Jamaica (2010): "the central bank lowered the policy rate to 6.25 percent..." with staff neutral: "Neutral[Staff did not challenge the authorities’ choices]" and future stance "Neutral[No mention on future monetary policy changes]".
  - NA / Not applicable where monetary policy is not independent (ECCU members, dollarized economies).
- Instruments used (explicit examples):
  - Interest rate decreases:
    - Haiti (2004): "During August–October 2004, interest rates were reduced to 7.6 percent from 13.6 percent..." (listed as "Decreased interest rates").
    - Jamaica (2010): "the central bank lowered the policy rate to 6.25 percent..."
    - Samoa: "Decrease interest rates" — "lower interest rates..."
  - Interest rate increases:
    - Moldova (2007): "raised policy interest rates by 2.5 percentage points..."
    - Sri Lanka (2016): "CBSL raised the policy rate by 25 basis points..."
    - Jamaica (2004, 2007): "moderate increases in domestic interest rates..."
  - Broad money control:
    - Haiti (1998): "Decreased broad money growth rate" — "use open market operations to control the money supply..."
- IMF agreement with authorities:
  - Frequently Y (agreement): e.g., Fiji — "Maintaining an accommodative monetary policy stance was appropriate..."; Haiti (1998) — "Directors welcomed the authorities’ intention..."; Moldova (2000, 2007) — "Y...Monetary policy was successfully tightened..."/"agreed that monetary policy should remain tight..."
  - Occasionally N (disagreement or concern): Haiti (2004) — "Directors expressed concern..." about interest rate reductions; Haiti (2004) IMF advice: "Tighten...Monetary policy needs to be tightened..."
  - Neutral or no challenge in some cases: Jamaica (2010) — "Neutral[Staff did not challenge the authorities’ choices]".

### Representative exact figures and passages preserved verbatim from excerpts
- "3.8 percent in 2017 from 0.4 percent in 2016..."
- Grenada (2004): "The economy contracted by 3 percent in 2004..."
- Haiti (2004): "The property damage and the interruption to economic activity are estimated to have totaled 5½ percent of GDP..."
- Haiti (1998): "the gourde appreciated by about 3½ percent with respect to the U.S. dollar and by about 9½ percent in real effective terms..."
- Haiti (2004): "During August–October 2004, interest rates were reduced to 7.6 percent from 13.6 percent..."
- Jamaica (2010): "the central bank lowered the policy rate to 6.25 percent..."
- Micronesia (2015): "The ratio of international reserves to money base was 96 percent at end-2015, compared with the statutory mandate of 60 percent..."
- Moldova (2007): "and raised policy interest rates by 2.5 percentage points..."
- Sri Lanka (2016): "the Central Bank of Sri Lanka (CBSL) raised the policy rate by 25 basis points..."

*Excerpted and summarized from Working Paper No. WP/2022/067 — Monetary Policy in Disaster-Prone Developing Countries (selected country Q&A excerpts).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022067-print-pdf.pdf_