## _wp1207

## Source details

**Canonical URL:** [_wp1207](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1207.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1207.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1207.pdf.json)

---

### Definition of Shock Episodes
- Purpose: Provide a historical perspective on the role of international reserves as a buffer against large external shocks in low-income countries (LICs) over the last three decades and assess whether macroeconomic costs associated with external shocks were larger in LICs with lower international reserve holdings prior to a shock.
- Main sample period: 1980–2007; crisis period analyzed separately: 2008–2010 (four-year event window centered in 2008).
- Outcomes measured: Forgone growth of real GDP and consumption per capita; additional macro variables during the crisis include real per-capita investment and external current account.
- Operational definition of an external shock event:
  - A “significant worsening” in climatic conditions, terms of trade, external demand, FDI, or aid inflows.
  - Annual percentage change of the shock variable falls below the bottom 10th percentile of its country-specific distribution (bottom 10th percentile obtained excluding the 1st and 99th percentiles).
- Event-window and cost measurement:
  - Five-year event window from one year before the shock (T-1) to three years after the shock (T+3).
  - Economic cost = cumulated sum of estimated growth losses relative to pre-shock trends over years following the shock.
  - Pre-shock trend = average growth rate in the three years preceding the shock.
  - Growth loss = negative difference between “shock” growth index (actual growth) and “no-shock” growth index (trend), both set to 100 in T-1.
- Assumptions and simplifications:
  - Country-specific shock distributions treated as independent across shocks and countries.
  - Contiguous shock years or shock years less than three years apart treated as a single episode occurring in the earliest year.
  - Combined-shock episodes not explicitly controlled for; combined shocks were relatively infrequent (one out of six shock years had two simultaneous shocks; one out of thirty had three simultaneous shocks).
  - Assumes same frequency of shocks for each country.

### Anatomy of Shocks: Frequency and Size (1980–2007)
- Typical frequency:
  - Between 1980 and 2007 LICs faced a large shock event every ten years on average.
  - By construction, a shock corresponds to roughly 10 percent of country observations.
- Likelihoods by type (text summary):
  - FDI shocks: about 16 percent.
  - Climatic shocks: about 13 percent.
  - Terms-of-trade shocks: about 11 percent.
  - External demand shocks: less frequent (relative statement).
- Typical sizes:
  - External demand and terms-of-trade shocks: about two standard deviations from the sample mean.
  - Aid shock: decline of 5 percentage points of GDP.
  - FDI shock: decline of 3 percentage points of GDP.
  - Example: A terms-of-trade shock implies a fall in relative prices of about 29 percent.

### Macroeconomic Impact and Cost of Shocks (event-study, 1980–2007)
- Event-window median behavior:
  - Median GDP growth decelerated after all shock types, particularly after terms-of-trade and external demand shocks.
  - Median real per-capita consumption growth remained subdued over the event window, especially following external demand shocks.
  - Current account deficit generally widened and remained depressed for about two years before returning to pre-shock levels.
  - Reserve coverage (months of imports):
    - Fell following shocks to external demand and terms of trade.
    - Increased steadily after shocks to FDI and aid flows.
- Representative quantitative outcomes (preserved exact figures from source):
  - Shock definition percentile: bottom 10th percentile (country-specific distribution, excluding 1st and 99th percentiles).
  - Analysis window: five-year event window (T-1 to T+3).
  - Pre-shock trend: average growth in the three years preceding the shock.
  - Sample period: 1980–2007; crisis period 2008–2010 assessed separately (four-year window centered in 2008).
  - Large shock event roughly every ten years on average (10 percent by construction).
  - Likelihoods: FDI about 16 percent; climatic 13 percent; terms of trade 11 percent.
  - Magnitudes: external demand and terms of trade ~ two standard deviations; aid shock ~ decline of 5 percentage points of GDP; FDI shock ~ decline of 3 percentage points of GDP.
  - Terms-of-trade shock example: fall in relative prices of about 29 percent.

### Macroeconomic Costs by Shock Type (key quantified findings)
- External-demand shock:
  - Cumulative losses as forgone GDP growth as high as 2 percentage points of GDP over four years.
  - Forgone real per-capita consumption growth: 7 percentage points over six years.
- Worsening in FDI inflows:
  - Typical annual GDP growth cost: approximately 0.2 percentage points.
  - Cumulative consumption losses: modest and short-lived.
- Climatic conditions shock:
  - Typical annual GDP growth cost: approximately 0.2 percentage points.
  - In island economies, cumulative GDP losses reached approximately 14 percentage points of GDP after a decade.
- Terms-of-trade shock:
  - Not accompanied by output loss (GDP growth).
  - Forgone real per-capita consumption growth: cumulated loss of 18 percentage points over thirteen years.
  - In islands and highly-indebted economies, consumption costs exceeded 20 percentage points in ten years.

### Structural Vulnerabilities and Heterogeneity of Impact
- Structural characteristics considered: exchange rate regime (Ghosh et al., 2002 classification), export/import concentration (commodity exporter/importer), indebtedness (highly-indebted if average debt-to-GDP > 50 percent), presence of an IMF program (assessed T-1 to T+3).
- Key differentiated impacts:
  - External demand shocks: larger impact in highly-indebted and commodity-exporting economies.
  - Terms-of-trade shocks: countries with a fixed exchange rate regime showed a protracted slowdown and decline in per-capita consumption.
  - Climatic shocks: more visible consequences in islands and commodity-exporting countries.
  - FDI and aid shocks: more prolonged impact in countries without an IMF program around the shock.

### The Role of International Reserves
- Reserve coverage grouping: reserves in the year preceding the shock classified as strictly above or below three months of imports.
- General findings:
  - Under most shocks, annual losses declined as reserve holdings approached three months of imports.
  - Countries with reserve coverage above three months of imports did not suffer any GDP growth losses except in the case of shocks to FDI inflows.
  - Economies with less than three months of imports generally encountered significant costs—particularly under an external demand shock with cumulative losses reaching approximately 13 percentage points over eight years.
- Consumption smoothing and reserves:
  - Higher reserves associated with a smoother adjustment of real per-capita consumption growth.
  - Except for FDI shocks, costs in terms of forgone consumption growth were generally about 2 percentage points in countries with more than three months of imports.
  - Countries with lower reserve coverage presented an average loss of about 1 percentage point for nearly 6 years.
  - For terms-of-trade shocks, cumulative losses were more than twenty times bigger in low-reserve economies than in economies with more than three months of imports.
- Caveat: Data limitations prevented an assessment of statistical significance for reserve comparisons.

### Structural Modulation of Reserve Benefits
- Reserve coverage above three months not uniformly protective:
  - Not a safe haven for GDP losses in island economies and countries with an exchange rate peg facing a terms-of-trade shock.
  - Significant cumulative losses in oil-exporting and commodity-importing countries after a significant worsening of external demand.
- Consumption outcomes and exchange regime:
  - Under a fixed exchange regime, consumption costs were somewhat smaller in countries with more than three months of imports prior to the shock.
  - Among countries with reserves ≤ 3 months, losses were larger in commodity-importing countries, islands, and highly-indebted economies.
- Interaction with IMF programs:
  - Countries without an IMF program generally experienced a more prolonged adjustment than countries with an IMF program in proximity to a shock episode.
  - An early IMF program engagement near a shock was generally associated with very limited GDP losses, but did not produce tangible benefits in shielding consumption (possibly reflecting fiscal consolidation).

### Robustness Analysis
- Alternative shock definitions:
  - 25th percentile threshold:
    - Shock episodes more frequent, smaller magnitude; role of reserves broadly confirmed.
    - With 25th percentile threshold, countries with reserve coverage above three months did not experience any loss in forgone GDP growth after an external demand shock.
  - Whole-sample distribution (10th percentile of whole-sample):
    - Shocks less frequent but larger magnitude; results broadly in line with main analysis, including reserves as a buffer.
    - GDP growth costs nil for countries with reserves > three months.
- Closed vs open events:
  - “Closed” event: shock growth index catches up with no-shock index within sample.
  - “Open” event: shock index consistently above or below no-shock index.
  - Cumulative losses for “closed” events broadly consistent but generally smaller and shorter, suggesting permanent factors do not significantly bias reserves-as-buffer finding.
- Data limitations:
  - Constraints prevented analysis of climatic shocks in some robustness exercises and prevented statistical significance testing for reserve comparisons.

### The Global Financial Crisis (2008–2010): Impact and Reserves Role
- Median macroeconomic impact (2007–2009):
  - Median GDP growth dropped by 3 percentage points between 2007 and 2009.
  - Median real per-capita consumption dropped by more than 4 percentage points between 2007 and 2009.
  - In 2008, median current account deficit deteriorated by about 3 percentage points of GDP.
- Reserve coverage during crisis:
  - Median reserve coverage fell by almost one month of imports in 2008.
  - In 2009, median coverage increased by one and a half months of imports as countries adjusted and the SDR allocation from the Fund became available.
- Reserves as buffer during the crisis:
  - Countries entering the crisis with reserve coverage above three months of imports were better able to buffer economic activity and smooth consumption than economies with lower reserves in 2007.
  - Economies with fewer reserves showed a sharper correction in median real per-capita investment, suggesting investment cuts to protect consumption where reserves were slender.
- Crisis heterogeneity (selected):
  - Median GDP growth declined in island economies with reserves ≤ 3 months prior to the shock, while remaining largely positive in islands with reserves > 3 months.
  - Slow growth observed in countries with high debt, fairly concentrated exports, and no financial arrangement with the IMF.
  - Highly-indebted countries with reserves ≤ 3 months confirmed vulnerability in real per-capita consumption growth.
  - Higher reserve coverage could not prevent significant adjustment of consumption patterns in commodity-importing and island economies.

### Data, Sample, and Methods (Appendix)
- Sample: panel of 71 LICs comprising 17 island economies, 17 commodity exporting countries, 6 oil exporters, 20 commodity importers, and 64 highly-indebted economies.
- Sample periods:
  - Event study: 1980 to 2007.
  - Current crisis study: 2007 to 2010.
- Primary data for 1980–2007 from April 2010 WEO; crisis period updated with September 2011 WEO.
- Historical per-capita consumption reconstructed using growth rates from Penn World Tables (version 6.3).
- Computations performed with Stata® 11/SE; code available on request.
- Key variable sources preserved exactly as in the source:
  - Exchange rate regime — Ghosh, Gulde, and Wolf's Database (2002).
  - IMF program — MONA.
  - Current account, Imports, International reserves, Real GDP growth, Real per-capita consumption growth, Real per-capita investment growth, Real per-capita absorption growth — World Economic Outlook Database.
  - Total debt — Lane and Milesi-Ferretti's Database (2007).
  - Aid flows — OECDstat Database.
  - Climatic conditions — OFDA/CRED International Disaster Database.
  - External demand — Global Economic Environment Database (IMF internal).
  - FDI — World Economic Outlook Database.
  - Terms of trade (goods only) — World Economic Outlook Database.

### Representative Summary Findings and Policy Implications
- Reserve adequacy matters: reserves above three months of imports in the year prior to a shock were generally associated with smaller GDP growth losses and smoother consumption adjustments, but three months is at most a lower bound.
- Structural characteristics critically shape shock impact and reserve effectiveness: exchange rate flexibility, debt levels, export concentration, and IMF program presence materially modify outcomes.
- Exchange rate flexibility functions as a shock absorber: countries with a flexible exchange rate prior to shock showed hardly any GDP loss in many cases.
- IMF engagement can limit GDP losses but may not shield consumption: early IMF program engagement near shocks generally associated with limited GDP losses; consumption protection less evident, possibly due to required fiscal consolidation.
- Optimal reserve policy should be country-specific: the appropriate level of reserve holdings varies with structural characteristics and shock exposures; the standard three-months-of-imports rule is a lower bound, not a universal target.

*Source: April 2010 WEO; September 2011 WEO; IMF working paper content unit.*

### 1.     Definition of Shock Episodes ....................................................................................

### 1.     Definition of Shock Episodes

### Introduction
- Purpose: Provide a historical perspective on the role of international reserves as a buffer against large external shocks in low-income countries (LICs) over the last three decades and assess whether macroeconomic costs associated with external shocks were larger in LICs with lower international reserve holdings prior to a shock.
- Key questions addressed:
  - What are the relevant external shock episodes in LICs?
  - What are the macroeconomic costs associated with them?
  - Do such costs vary with structural characteristics of the economy, including the level of international reserves?
- Sample and periods:
  - Main sample period: 1980–2007.
  - Crisis period analyzed separately: 2008–2010 (four-year event window centered in 2008).
- Outcomes measured: Forgone growth of real GDP and consumption per capita; additional macro variables during the crisis include real per-capita investment and external current account.

### Methodology
- Definition of an external shock event:
  - A “significant worsening” in a country’s climatic conditions, terms of trade, external demand, FDI, or aid inflows.
  - Operational criterion: annual percentage change of the shock variable falls below the bottom 10th percentile of its country-specific distribution.
  - Bottom 10th percentile threshold obtained from the country-specific distribution by excluding the 1st and 99th percentiles.
- Event-window and cost measurement:
  - Five-year event window capturing behavior from one year before the shock episode (T-1) to three years after the shock (T+3).
  - Economic cost defined as the cumulated sum of estimated growth losses with respect to pre-shock trends over years following the shock episode.
  - Pre-shock trend: average growth rate of the relevant variable in the three years preceding the shock.
  - Growth loss computed as negative difference between the “shock” growth index (actual growth) and the “no-shock” growth index (trend growth), with both indices set to 100 in the year preceding the shock (T-1).
  - Number of years with a loss interpreted as duration of the economic cost.
- Assumptions and simplifications:
  - Country-specific distributions of shock variables are treated as independent across shocks and countries.
  - Contiguous shock years or shock years less than three years apart are treated as a single episode occurring in the earliest relevant year.
  - Combined-shock episodes (simultaneous shocks of different nature) are not controlled for explicitly in cost assessment; combined shocks were relatively infrequent (one out of six shock years had two simultaneous shocks; one out of thirty had three simultaneous shocks).
  - Assumes same frequency of shocks for each country (implies any country is equally likely to face a shock).
- Advantages of this shock definition:
  - Controls for heterogeneity among LICs using country-specific distributions.
  - Captures rare events by focusing on bottom 10th percentile.
  - Centers analysis on reaction to shock by assuming same shock frequency across countries.

### Anatomy of Shocks: Frequency and Size (1980–2007)
- Typical frequency:
  - Between 1980 and 2007 LICs were confronted with a large shock event every ten years on average.
  - A shock corresponds to roughly 10 percent of country observations by construction.
- Likelihoods by type (text summary):
  - FDI shocks: about 16 percent.
  - Climatic shocks: about 13 percent.
  - Terms-of-trade shocks: about 11 percent.
  - External demand shocks: less frequent (relative statement).
- Typical sizes:
  - External demand and terms-of-trade shocks: about two standard deviations from the sample mean.
  - Aid and FDI shocks: about one standard deviation; quantified as declines of 5 and 3 percentage points of GDP respectively.
  - Example: A terms-of-trade shock implies a fall in relative prices of about 29 percent (as stated in the source).

### Macroeconomic Impact and Cost of Shocks
- Event-window behavior (median values used to limit outlier impact):
  - Median GDP growth rate decelerated in the aftermath of all shock types, particularly after terms-of-trade and external demand shocks.
  - Median real per-capita consumption growth remained subdued over the event window, especially following external demand shocks.
  - Current account deficit generally widened and remained depressed for about two years before returning to pre-shock levels.
  - Reserve coverage (months of imports) trends:
    - Fell following shocks to external demand and terms of trade, suggesting use of international reserves to finance imports when foreign exchange is constrained.
    - Increased steadily after shocks to FDI and aid flows, suggesting foreign financing flows in LICs have large import content.
- Measurement notes:
  - The cumulated sum of annual losses (in percentage points) represents the cost associated with a shock; the example figure illustrates a duration of seven years and a negative cumulative cost (loss).

### Structural Vulnerabilities and Heterogeneity of Impact
- Structural characteristics considered:
  - Exchange rate regime (classification follows Ghosh et al., 2002; regime prevailing in the year before the shock).
  - Export and import concentration (commodity exporter/importer classification).
  - Level of indebtedness (highly-indebted economies defined as average debt-to-GDP ratio higher than 50 percent over the sample period).
  - Presence of an IMF program (assessed during the period from one year prior to the shock to three years after).
- Key differentiated impacts:
  - External demand shocks:
    - Larger macroeconomic impact in highly-indebted and commodity-exporting economies, reflecting greater dependence on exports as a source of financing.
  - Terms-of-trade shocks:
    - Countries with a fixed exchange rate regime showed a protracted slowdown in economic activity and a decline in per-capita consumption.
  - Climatic shocks:
    - More visible macroeconomic consequences in island and commodity-exporting countries, possibly due to concentrated production bases.
  - FDI and aid shocks:
    - Significant worsening in FDI and aid flows had a fairly prolonged impact in countries without an IMF program around the time of the shock episode.

### Summary of Key Quantitative Points (explicit figures preserved)
- Shock definition percentile: bottom 10th percentile (country-specific distribution, excluding 1st and 99th percentiles).
- Analysis window: five-year event window (T-1 to T+3).
- Pre-shock trend definition: average growth in the three years preceding the shock.
- Sample period: 1980–2007 (event-study analysis); crisis period 2008–2010 assessed separately (four-year window centered in 2008).
- Representative frequencies and sizes:
  - Large shock event roughly every ten years on average (10 percent by construction).
  - Likelihoods: FDI about 16 percent; climatic 13 percent; terms of trade 11 percent.
  - Magnitudes: external demand and terms of trade ~ two standard deviations; aid shock ~ decline of 5 percentage points of GDP; FDI shock ~ decline of 3 percentage points of GDP.
  - Terms-of-trade shock example: fall in relative prices of about 29 percent.

*Source: April 2010 WEO, IMF.*

### 12.      The macroeconomic costs associated with external shocks were large and

### _wp1207 - 12.      The macroeconomic costs associated with external shocks were large and persistent

### Macroeconomic costs by type of external shock
- External-demand shock:
  - Cumulative losses expressed as forgone GDP growth were as high as 2 percentage points of GDP over four years.
  - In terms of real per-capita consumption growth, costs amounted to 7 percentage points over six years.
- Worsening in FDI inflows:
  - Macroeconomic costs were typically modest, approximately 0.2 percentage points of GDP growth in a year.
  - Cumulative consumption losses were modest and short-lived.
- Climatic conditions shock:
  - Macroeconomic costs were typically modest, approximately 0.2 percentage points of GDP growth in a year.
  - In island economies, cumulative GDP losses reached approximately 14 percentage points of GDP after a decade.
- Terms-of-trade shock:
  - Not accompanied by any output loss (GDP growth).
  - Costs in terms of forgone real per-capita consumption growth were large: consumption growth remained below trend for thirteen years resulting in a cumulated loss of 18 percentage points.
  - In islands and highly-indebted economies, costs associated with a significant worsening in the terms of trade exceeded 20 percentage points of consumption in ten years.

### Heterogeneity: role of structural characteristics
- Larger and more persistent GDP growth costs observed in:
  - Islands.
  - Commodity-exporting economies.
  - Commodity-importing economies.
- Exchange rate regime:
  - Countries with a flexible exchange rate regime prior to a shock event showed hardly any loss, highlighting exchange rate flexibility as a shock absorber.
  - Countries with a fixed exchange rate regime normally suffered larger consumption losses than economies with a flexible regime.
- IMF program engagement:
  - An early program engagement with the IMF in the proximity of a shock event was generally associated with very limited GDP losses.
  - Presence of an IMF-supported program did not produce tangible benefits in shielding consumption, possibly reflecting fiscal consolidation in such arrangements.
- Highly-indebted economies:
  - Reserves above three months of imports appeared to be an adequate cushion against all shocks in highly-indebted economies.

### The role of international reserves
- Reserve coverage grouping:
  - LICs were grouped by whether reserve coverage in the year preceding the shock was strictly above or below three months of imports.
- General findings:
  - Under most shocks, annual losses declined as reserve holdings approached three months of imports, suggesting the three-months rule is at most a lower bound for LICs.
  - Countries with reserve coverage above three months of imports did not suffer any GDP growth losses except in the case of shocks to FDI inflows.
  - Economies with less than three months of imports generally encountered significant costs—particularly under an external demand shock with cumulative losses reaching approximately 13 percentage points over eight years.
- Consumption smoothing:
  - Higher reserves were associated with a smoother adjustment of real per-capita consumption growth.
  - Except for FDI shocks, costs in terms of forgone consumption growth were generally about 2 percentage points in countries with more than three months of imports.
  - By contrast, countries with lower reserve coverage presented an average loss of about 1 percentage point for nearly 6 years.
  - In the case of a terms-of-trade shock, cumulative losses were more than twenty times bigger in low-reserve economies than in economies with more than three months of imports.
- Caveat:
  - Comparison of costs across reserve coverage groups warrants caution as data limitations prevented an assessment of statistical significance.

### Structural modulation of reserve benefits
- Reserve coverage above three months was not uniformly protective:
  - Not a safe haven for GDP losses in island economies and countries with an exchange rate peg facing a terms-of-trade shock.
  - Cumulative losses were significant in oil-exporting and commodity-importing countries after a significant worsening of external demand.
- Consumption outcomes:
  - Under a fixed exchange regime, consumption costs were somewhat smaller in countries with more than three months of imports prior to the shock episode.
  - Among countries with reserve coverage below three months of imports, losses were larger in commodity-importing countries, islands, and highly-indebted economies.
- Interaction with IMF programs:
  - Countries without an IMF program generally experienced a more prolonged adjustment than countries with an IMF program in proximity to a shock episode.

### Robustness analysis
- Alternative shock definitions examined:
  - Threshold: 25th percentile (broader range of events).
    - Results confirm previous findings: shock episodes more frequent, smaller magnitude; role of reserves broadly confirmed though magnitude and duration differed somewhat.
    - With 25th percentile threshold, countries with reserve coverage above three months of imports prior to the shock did not experience any loss in terms of forgone GDP growth after an external demand shock.
  - Distribution: whole-sample distribution (10th percentile of whole-sample).
    - Shocks less frequent but of larger magnitude; results broadly in line with prior analysis, including reserves as a buffer.
    - Costs in terms of GDP growth were nil for countries with reserve coverage above three months of imports prior to the shock.
- Closed vs open events:
  - “Closed” event: country’s “shock” growth index catches up with its “no-shock” index within the sample period.
  - “Open” event: “shock” index is consistently above or below the “no-shock” index.
  - Cumulative losses for “closed” events broadly consistent with main results but generally smaller in magnitude and shorter in duration, suggesting permanent factors do not significantly bias the reserves-as-buffer finding.
- Data limitations:
  - Data constraints prevented the analysis of shocks to climatic conditions in some robustness exercises and prevented statistical significance assessment for reserve comparisons.

### The global financial crisis (2008–2010): impact and reserves role
- Macroeconomic impact (median values, 2007–2009):
  - Median GDP growth dropped by 3 percentage points between 2007 and 2009.
  - Median real per-capita consumption dropped by more than 4 percentage points between 2007 and 2009.
  - In 2008, the median current account deficit deteriorated by about 3 percentage points of GDP.
- Reserve coverage during crisis:
  - Median reserve coverage fell by almost one month of imports in 2008.
  - In 2009, median coverage increased by one and a half months of imports as countries adjusted and the SDR allocation from the Fund became available.
- Reserves as buffer during the crisis:
  - Countries entering the crisis with reserve coverage above three months of imports were better able to buffer economic activity and smooth consumption than economies with lower reserves in 2007.
  - Economies with fewer reserves showed a sharper correction in median real per-capita investment, suggesting investment cuts to protect consumption where reserves were slender.
- Heterogeneity during the crisis:
  - Median GDP growth declined in island economies with reserve coverage below three months of imports prior to the shock, while remaining largely positive in islands with higher reserves.
  - Slow growth observed in countries with high debt, fairly concentrated exports, and no financial arrangement with the IMF.
  - Highly-indebted countries with reserve below three months of imports confirmed vulnerability in real per-capita consumption growth.
  - Higher reserve coverage could not prevent significant adjustment of consumption patterns in commodity-importing and island economies.

*Source: April 2010 WEO; September 2011 WEO, IMF (as presented in the content unit).*

### 25.      This paper investigates the issue of reserve adequacy in LICs. Specifically, it

### 25.      This paper investigates the issue of reserve adequacy in LICs. Specifically, it examines the role of international reserves as a buffer against different types of shocks (i.e. external-demand, terms-of-trade, climatic, FDI, and aid shocks) over the last three decades.

### Objective and scope
- Examines role of international reserves as a buffer against different types of shocks: external-demand, terms-of-trade, climatic, FDI, and aid shocks.
- Expands previous research by:
  - Examining a wide variety of external shocks.
  - Explicitly taking into account LICs’ different structural characteristics: exchange rate regime, export and import concentration, level of indebtedness, and presence of an IMF program.
- Period covered for the event study analysis: 1980–2007.

### Event-study findings (1980–2007)
- LICs faced large external shocks; macroeconomic costs varied by:
  - Nature of the shock.
  - Structural characteristics of the economy.
  - Level of international reserve holdings.
- Over the three years following any shock event:
  - Macroeconomic performance weakened significantly.
  - Median real GDP growth and external accounts deteriorated—particularly in countries with:
    - A high level of debt.
    - A fairly concentrated export base.
    - Limited exchange rate flexibility.
- Costs relative to pre-shock trends were considerable in terms of:
  - Forgone real GDP.
  - Per-capita consumption growth.
- After a terms-of-trade shock:
  - Costs were rather large and persistent in commodity-exporting economies (including oil-exporters), islands, and countries without an IMF program in the proximity of the shock event.
- Reserve coverage and cushioning effects:
  - Countries with international reserve holdings above three months of imports in the year before the shock event were generally able to better cushion economic activity.
  - Higher reserve coverage was not a panacea—especially in countries with an exchange rate peg.
  - The standard rule of thumb of three months of imports is at most a lower bound; the appropriate level of reserve holdings may vary with the structural characteristics of the economy.

### Role of international reserves during the global crisis (2008–2010)
- International reserves helped LICs withstand the impact of the current global crisis (2008-2010).
- LICs with reserve coverage above three months of imports in 2007 were able to better cushion GDP growth and consumption from crisis spillovers.
- The potential benefits of holding international reserves varied markedly depending on the structural characteristics of LICs.

### Implications and conclusions
- Reserve adequacy matters for LICs’ ability to absorb diverse external shocks.
- Structural characteristics (exchange rate regime, debt level, export concentration, IMF program presence) critically shape both shock impact and the effectiveness of reserves.
- The three months of imports benchmark is a lower bound; optimal reserve levels should be tailored to country-specific structural features and shock exposures.

*Source: IMF working paper content unit.*

### 8.      IMF (2011); Managing Volatility: A Vulnerability Exercise for Low- Income

### _wp1207 - 8.      IMF (2011); Managing Volatility: A Vulnerability Exercise for Low- Income

### Appendix — data, sample, and methods
- Sample: panel of 71 LICs comprising 17 island economies, 17 commodity exporting countries, 6 oil exporters, 20 commodity importers, and 64 highly-indebted economies (Table A.1).
- Sample periods:
  - Event study analysis: 1980 to 2007.
  - Study on the impact of the current crisis: 2007 to 2010.
- Data sources and notes:
  - Primary data for 1980–2007 from April 2010 WEO database; a few exceptions documented in Table A.1.
  - Historical per-capita consumption reconstructed using growth rates from Penn World Tables (version 6.3) to reduce missing observations.
  - Current crisis period updated with September 2011 WEO database.
  - Computations performed with Stata® 11/SE; code available on request.

### Variables and data sources (Table A.1)
- Economic variables (unit — database):
  - Exchange rate regime — index — Ghosh, Gulde, and Wolf's Database (2002).
  - IMF program — index — MONA.
  - Current account — USD billions — World Economic Outlook Database.
  - Imports — USD billions — World Economic Outlook Database.
  - International reserves — USD billions — World Economic Outlook Database.
  - Real GDP growth — percent change — World Economic Outlook Database.
  - Real per-capita consumption growth — percent change — World Economic Outlook Database.
  - Real per-capita investment growth — percent change — World Economic Outlook Database.
  - Real per-capita absorption growth — percent change — World Economic Outlook Database.
  - Total debt — in percent of GDP — Lane and Milesi-Ferretti's Database (2007).
- Shock variables (unit — database):
  - Aid flows (excluding debt relief) — USD millions — OECDstat Database.
  - Climatic conditions — index — OFDA/CRED International Disaster Database.
  - External demand — percent change — Global Economic Environment Database (IMF internal).
  - FDI — USD billions — World Economic Outlook Database.
  - Terms of trade (goods only) — index — World Economic Outlook Database.

### Sample listing and structural characteristics (Table A.2)
- Countries listed with indicators for: Islands (x), Commodity exporters (x), Oil exporters (x), Commodity importers (x), Highly-indebted (x).
- Note definitions:
  - Commodity exporters: as defined by April 2010 WEO.
  - Commodity importers: countries with share of commodity imports to total merchandise imports exceeding 40 percent on average over sample.
  - Highly-indebted: countries with average debt-to-GDP ratio exceeding 50 percent over sample.

### Shock variables — summary statistics (1980–2007) (Table A.3; average percent changes)
- Memorandum averages:
  - LICs: External demand 3.8; Terms of trade (goods only) -0.8; FDI 20.9; Aid excluding debt relief 8.1.
  - East Asia & Pacific: External demand 4.8; Terms of trade 0.3; FDI 11.0; Aid 8.4.
  - Europe & Central Asia: External demand 2.6; Terms of trade -1.9; FDI 33.1; Aid 26.3.
  - Latin America & Caribbean: External demand 3.1; Terms of trade -0.7; FDI 13.0; Aid 5.1.
  - Middle East & North Africa: External demand 5.2; Terms of trade -2.1; FDI 19.7; Aid 4.4.
  - South Asia: External demand 4.1; Terms of trade 0.1; FDI 16.1; Aid 10.6.
  - Sub-Saharan Africa: External demand 3.7; Terms of trade -1.1; FDI 24.0; Aid 5.6.
- Country examples (average percent changes shown exactly as in table, selected):
  - Afghanistan: External demand 3.8; Terms of trade 0.0; FDI 31.6; Aid 24.7.
  - Bangladesh: External demand 2.7; Terms of trade -0.9; FDI 21.8; Aid -0.6.
  - Chad: External demand 3.2; Terms of trade 3.8; FDI 7.0; Aid 5.4.
  - Mongolia: External demand 7.6; Terms of trade 1.4; FDI 34.6; Aid 30.7.
  - Timor-Leste: External demand n.a.; Terms of trade 1.2; FDI -7.5; Aid 33.1.
  - Nigeria: External demand 3.1; Terms of trade 0.6; FDI 13.6; Aid 17.4.

### Macroeconomic variables — summary statistics (1980–2007) (Table A.4; percent changes unless otherwise indicated; average values)
- Memorandum averages:
  - LICs: Real GDP 3.0; Real per-capita absorption 1.2; Real per-capita consumption 1.5; Real per-capita investment 2.1; Reserves in months of imports 3.1; Reserve accumulation 12.6; CA-to-GDP (in %) -6.1.
  - East Asia & Pacific: Real GDP 4.0; Real per-capita absorption 1.9; Real per-capita consumption 1.7; Real per-capita investment 3.0; Reserves in months 3.1; Reserve accumulation 19.7; CA-to-GDP -3.5.
  - Europe & Central Asia: Real GDP 0.3; Real per-capita absorption 6.5; Real per-capita consumption 10.3; Real per-capita investment 7.3; Reserves in months 2.9; Reserve accumulation 30.2; CA-to-GDP -5.3.
  - Latin America & Caribbean: Real GDP 2.6; Real per-capita absorption 0.9; Real per-capita consumption 0.9; Real per-capita investment 1.4; Reserves in months 2.4; Reserve accumulation 8.7; CA-to-GDP -12.5.
  - Middle East & North Africa: Real GDP 2.7; Real per-capita absorption -0.3; Real per-capita consumption -1.1; Real per-capita investment 4.8; Reserves in months 5.4; Reserve accumulation 9.4; CA-to-GDP -1.7.
  - South Asia: Real GDP 7.1; Real per-capita absorption 3.6; Real per-capita consumption 3.2; Real per-capita investment 4.4; Reserves in months 4.9; Reserve accumulation 17.8; CA-to-GDP -6.4.
  - Sub-Saharan Africa: Real GDP 2.7; Real per-capita absorption 0.0; Real per-capita consumption 0.0; Real per-capita investment 0.6; Reserves in months 2.9; Reserve accumulation 8.1; CA-to-GDP -8.0.
- Country examples (selected, averages preserved as in table):
  - Afghanistan: Real GDP 11.7; Real per-capita absorption 1.6; Real per-capita consumption 2.0; Real per-capita investment -0.4; Reserves in months of imports 3.5; Reserve accumulation 36.4; CA-to-GDP -5.3.
  - Bangladesh: Real GDP 4.4; Real per-capita absorption 1.8; Real per-capita consumption 1.0; Real per-capita investment 5.0; Reserves in months 2.5; Reserve accumulation 9.3; CA-to-GDP -1.6.
  - Liberia: Real GDP -1.0; Real per-capita absorption -5.6; Real per-capita consumption -5.2; Real per-capita investment -7.8; Reserves in months 0.2; Reserve accumulation 2.8; CA-to-GDP -26.7.
  - Timor-Leste: Real GDP 3.0; Reserves in months n.a.; Reserve accumulation 4.0; CA-to-GDP 33.3.
  - Uganda: Real GDP 5.1; Real per-capita absorption 4.5; Real per-capita consumption 3.6; Real per-capita investment 13.9; Reserves in months 4.1; Reserve accumulation 16.9; CA-to-GDP -3.7.

### Macroeconomic impact of shocks — median values (Table A.5 and A.6; annual percent changes unless otherwise indicated)
- Whole sample medians (selected series, T-1 to T+3):
  - GDP growth: T-1 3.4; T0 3.2; T+1 3.1; T+2 3.5; T+3 3.7 (first row of Table A.5 whole-sample block).
  - Real per-capita consumption growth: T-1 -1.2; T0 0.4; T+1 0.3; T+2 -0.1; T+3 0.8.
  - Real per-capita investment growth: T-1 -0.6; T0 2.4; T+1 -2.1; T+2 1.0; T+3 1.5.
  - Reserves (in months of imports): T-1 2.5; T0 2.3; T+1 2.2; T+2 2.4; T+3 2.5.
  - Reserves (percentage change): T-1 9.6; T0 5.2; T+1 4.8; T+2 9.9; T+3 8.9.
  - Current account (in % of GDP): T-1 -5.3; T0 -6.5; T+1 -5.9; T+2 -6.0; T+3 -4.5.
- Subsamples and shock-type blocks present median trajectories for:
  - Flexible exchange rate vs Fixed exchange rate.
  - Commodity exporters and Oil exporters.
  - Shock types: External Demand; Terms of Trade; FDI; Climatic Conditions; Aid.
- Table A.6 continues medians for additional subsamples:
  - No IMF program vs Early IMF program.
  - Commodity importers; Islands; Highly-indebted.
  - Reserves ≤ 3 months vs Reserves > 3 months.
- Exact numerical medians across T-1 to T+3 are reported in the tables for a large set of combinations (see Tables A.5 and A.6 for the full matrix of median outcomes by shock type and subsample).

### Figures — GDP and consumption costs by shock type, exchange regime, reserves, and subsample (Figures A.1–A.6)
- Figures present losses computed with respect to pre-shock trend in percentage points; median values; duration up to 10 years (T-1 through T+10).
- Panels include:
  - Whole sample; Flexible exchange rate; Fixed exchange rate; Commodity exporters; Oil exporters (Figures A.1, A.3, A.5).
  - Islands; Highly-indebted; No IMF program; Early IMF program; Commodity importers (Figures A.2, A.4, A.6).
  - Comparisons by reserves level: Reserves ≤ 3 months vs Reserves > 3 months (Figures A.3–A.6).
- Shock types shown in figures: External Demand; Terms of Trade; FDI; Climatic Conditions; Aid.
- Figures display median trajectories for:
  - GDP growth losses and per-capita consumption growth losses relative to pre-shock trend (scale and magnitudes visible across panels, including large negative peaks for some subsamples and shock types).

### SDR allocation (Table A.7; in millions of SDRs)
- Selected allocations (exact values preserved):
  - Afghanistan 128.6
  - Armenia 88
  - Bangladesh 463.3
  - Benin 49.8
  - Bhutan 6
  - Bolivia 137.4
  - Burkina Faso 48.2
  - Burundi 60.2
  - Cambodia 68.5
  - Cameroon 152.8
  - Cape Verde 8.5
  - Central African Republic 44
  - Chad 44.2
  - Comoros 7.8
  - Congo, Dem. Republic of 424.5
  - Congo, Republic of 70
  - Cote d'Ivoire 273.1
  - Djibouti 14
  - Dominica 7.2
  - Eritrea 15.2
  - Ethiopia 116.8
  - Gambia, The 24.6
  - Georgia 144
  - Ghana 290.9
  - Grenada 10.2
  - Guinea 84.9
  - Guinea-Bissau 12.4
  - Guyana 72.6
  - Haiti 64.8
  - Honduras 104.8
  - Kenya 222.7
  - Kiribati 5.3
  - Kyrgyz 84.7
  - Lao, People's Dem. Republic 41.3
  - Lesotho 29.1
  - Liberia 103
  - Madagascar 97.8
  - Malawi 55.4
  - Maldives 7.4
  - Mali 73.5
  - Mauritania 51.9
  - Moldova 117.7
  - Mongolia 48.8
  - Mozambique 108.8
  - Myanmar 202.3
  - Nepal 60
  - Nicaragua 105.1
  - Niger 53.5
  - Nigeria 1,518.20
  - Papua New Guinea 116.2
  - Rwanda 63.1
  - Samoa 9.9
  - São Tomé and Principe 6.5
  - Senegal 130.3
  - Sierra Leone 82.1
  - Solomon Islands 9.3
  - St. Lucia 13.8
  - St. Vincent and the Grenadines 7.6
  - Sudan 141.9
  - Tajikistan 82.1
  - Tanzania 159.1
  - Timor-Leste 7.7
  - Togo 59.4
  - Tonga 6.6
  - Uganda 143.7
  - Uzbekistan 262.8
  - Vanuatu 16.3
  - Vietnam 267.1
  - Yemen, Republic of 203.5
  - Zambia 400.8
  - Zimbabwe 328.4
- Note: * Countries marked with an asterisk will receive allocations for the first time as a result of both the General and Special SDR Allocation.

### Macroeconomic impact of the crisis — median values (Table A.8; annual percent changes unless otherwise indicated)
- Pooled medians for 2007–2010 (selected series and groupings preserved as in table):
  - Whole sample pooled medians (selected):
    - GDP growth: 2007 5.8; 2008 5.4; 2009 3.3; 2010 4.9.
    - Real per-capita consumption growth: 2007 4.1; 2008 4.2; 2009 0.7; 2010 2.2.
    - Real per-capita investment growth: 2007 5.0; 2008 5.0; 2009 3.0; 2010 5.7.
    - Reserves (in months of imports): 2007 4.0; 2008 3.4; 2009 5.0; 2010 4.7.
    - Reserves (percentage change): 2007 28.2; 2008 4.0; 2009 22.0; 2010 5.9.
    - Current account (in % of GDP): 2007 -8.0; 2008 -11.1; 2009 -8.5; 2010 -8.0.
  - Subsample and region-specific median trajectories are presented for:
    - Early IMF program; Oil exporters; Commodity importers; Islands; Highly indebted; No IMF program.
    - Reserves ≤ 3 Months of Imports vs Reserves > 3 Months of Imports.
- Data limitations: analysis not possible for LICs with different exchange rate regimes (footnote 1 in Table A.8).
- Source: September 2011 WEO, IMF.

*Source: IMF (April 2010 WEO; September 2011 WEO; Tables and Figures A.1–A.8 as provided in the source PDF).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1207.pdf_