## _wp12223 - Appendix I. Country Classifications and Data

## Source details

**Canonical URL:** [_wp12223 - Appendix I. Country Classifications and Data](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12223.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12223.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12223.pdf.json)

---

### I. Introduction — purpose, scope, core findings
- Purpose: consider features of small island states in the Asia-Pacific region and document key characteristics that set them apart from small states in other regions.
- Scope: restricted to macro-oriented indicators—population size, income per capita, fertility of land, ability to tap into economies of scale, and degree of geographic isolation.
- Exclusions: micro-oriented variables (telecommunications, electricity generation) and development indicators (literacy, infant mortality).
- Core finding preview:
  - Pacific small island states are typically extremely geographically isolated, have limited scope to tap into economies of scale due to small populations, face complications interpreting income per capita, and often lack fertile land constraining import substitution through agriculture.
- Organization: Sections II–VI discuss indicators; Section VII summarizes via ranking and relates to economic outcomes; Section VIII concludes.

### II. Defining small states by population
- Common metric: World Bank defines a small state as one with a population of less than 1½ million people (World Bank, 2007).
- Sample used: 50 small states (World Bank data).
  - Examples: Tuvalu with a population of only around 10,000 people; Trinidad and Tobago with around 1.3 million people.
- Heterogeneity note:
  - Population alone may not predict income; example: Monaco’s population is around three times Tuvalu’s, but GNI per capita is almost 40 times larger.
- Data exclusions: 15 territories excluded from initial World Bank sample of 65 due to unavailable comparable data.

### III. Lacking economies of scale
- Empirical relationship:
  - Population positively related to total income, but trend has weak explanatory power (Figure 2, R² = 0.244). Outliers exist (e.g., Luxembourg).
- Constraints for Pacific small islands:
  - Do not share attributes of financial centers or trade hubs (proximity to large skilled labor markets, infrastructure sharing, strategic shipping routes).
  - Geography precludes sharing infrastructure; duplication required for dispersed populations.
  - Some financial-service activities less tied to location (business registrations, ship registration) remain feasible.
- A. Fixed costs spread over narrow base
  - Small populations imply small total income and narrow tax base; fixed costs for public goods/services can be an unusually high share of national income.
  - Consequences: insufficient tax revenue for basic health services, transportation, government administration; small administrations with limited capacity.
  - Disaster-resilient infrastructure increases fixed-cost burden.
- B. Scope for regional cooperation
  - Regional aggregation could create larger common markets or shared infrastructure.
  - Aggregate-population illustration: Pacific islands’ regional aggregate population is not clearly disadvantaged relative to other regions (Figure 3).
  - Barriers: cultural, language, distance, legal, political heterogeneity.
  - Existing regional mechanisms: Pacific Islands Forum; Forum Economic Ministers Meeting (FEMM); Pacific Plan; Pacific Agreement on Closer Economic Relations; Pacific Financial Technical Assistance Center (PFTAC).
- Key GNI deviation examples (median GNI of small state sample was US$1,719 million, 2010):
  - Tuvalu: Per cent below median = -97.3
  - Palau: Per cent below median = -92.2
  - Monaco: Per cent above median = 276.9
  - Luxembourg: Per cent above median = 2,170.6
  - Channel Islands: Per cent above median = 495.8
  - Note: Of Pacific small countries, only Timor-Leste, Macau, and Fiji had GNI larger than the median US$1.7 billion in 2010.

### IV. Income per capita — limitations for microstates
- Use and limitations:
  - Common proxy for poverty and aid eligibility (IDA eligibility, IMF PRGT linkage).
  - Extremely small populations can inflate income per capita and mask inadequate total income to fund domestic services.
  - Example: high per-capita income but expensive out-of-country medical services reduce disposable per-capita consumption and welfare.
  - Distortions increase as denominator approaches extremely small values.
- Distributional note:
  - Top decile of small states by GNI per capita is entirely European (Monaco, Liechtenstein, Luxembourg, Channel Islands, San Marino).
- Policy implication:
  - Multiple vulnerabilities (small population, geographic isolation, low land fertility) compound to reduce ability to overcome disadvantages.

### V. Scope for import substitution through agriculture
- Common feature: many small states have limited fertile/arable land despite sometimes large total landmass.
- Risk channel: limited arable land → heavy dependence on food imports → greater exposure to balance of payments crises when food prices rise.
- Specifics and extremes:
  - Five states in sample have no arable land at all: Macao, Greenland, Monaco, Gibraltar, Tuvalu.
  - Tuvalu: Arable land m2 per capita = 0.
  - Extremes in arable land per capita:
    - Guyana: Arable Land m2 per capita = 5,567
    - Bahrain: Arable Land m2 per capita = 8
  - Some Pacific islands (Fiji, Samoa, Tonga, Timor-Leste) are relatively fertile compared with many other small states.
- Regional patterns:
  - Pacific states generally more infertile than small countries in Europe or Latin America and the Caribbean, but not as infertile as many Middle East and North Africa states or Greenland.
- Trade as mitigation:
  - Trade with agriculturally productive partners could offset domestic constraints, but geographic isolation raises transport costs and limits market access.
  - Fishing rights and fisheries exports provide possible revenue sources for Pacific islands, but isolation remains a constraint.

### VI. Geographic isolation
- Defining feature: extreme remoteness from continents and often from neighboring countries.
- Implications:
  - Higher transport costs for trade and commerce as distance increases.
  - Increased vulnerability from import dependence (especially food) and the costliness of accessing external services and markets.
  - Tourism viability is conditional on specific comparative advantages; distance and cost disadvantages limit feasibility.
  - Dispersion across islands increases susceptibility to natural disasters (tsunamis, hurricanes) and complicates resilient infrastructure provision.
  - Environmental vulnerability: rising sea levels and global warming disproportionately affect some Pacific islands; similar environmental pressures affect Caribbean small states.
- Isolation index and distance metrics:
  - Appendix I references the isolation sub index (EVI-13) of the Environmental Vulnerability Index calculated by SOPAC and the UNEP.
  - Samoa, Marshall Islands, Tuvalu, Kiribati, and Tonga are among the most isolated states in the world; each is "more than 3,000 kilometers from the nearest continent, Australia."
  - Pacific islands’ average distance to the nearest continent is "more than four to five times" that of the average country in the Caribbean or Sub-Saharan Africa.
  - Mayer and Zignago (2011) and Chen et al, 2012 measures reinforce that weighting distances by GDP makes Pacific small states appear even worse off.

### VII. Vulnerability ranking methodology and selected outcomes
- Five indicators used to rank 50 small states:
  - population, arable land per capita, distance to nearest continent, total GNI (US dollar level), and GNI per capita.
- Ranking convention: higher rank number indicates greater relative vulnerability on that indicator (e.g., Tuvalu rank 50 on population = most vulnerable by population).
- Aggregate index: equally weighted average across the five individual indicator rankings.
- Caveats: mechanical indices cannot fully reflect complexities; equal weighting implies no reasoned preference function.
- Selected ranking outcomes (rank averages):
  - Most vulnerable (by index average):
    - Tuvalu: Rank average 46.2 (Population 50, Arable land 50, Distance 48, GNI 50, GNI/Capita 33)
    - Kiribati: Rank average 41.6 (36, 34, 47, 46, 45)
    - Marshall Islands: Rank average 40.6 (43, 28, 49, 48, 35)
    - Micronesia, Fed. Sts.: Rank average 40.0 (32, 38, 44, 45, 41)
    - Palau: Rank average 38.4 (49, 26, 39, 49, 29)
  - Least vulnerable (by index average):
    - Luxembourg: Rank average 8.0 (18, 11, 7, 1, 3)
    - Estonia: Rank average 6.0 (2, 2, 2, 5, 19)
    - Cyprus: Rank average 12.4 (7, 20, 21, 2, 12)
    - Montenegro: Rank average 12.6 (14, 3, 3, 16, 27)
    - Guyana and Trinidad and Tobago: Rank average 15.2 each (Guyana: 10, 1, 1, 24, 40; Trinidad and Tobago: 1, 37, 18, 3, 17)
- Author’s interpretation:
  - Tuvalu is the most vulnerable small state in the sample due to very small population, no arable land, high isolation, and small GNI, outweighing its relatively favorable income per capita ranking.
  - Estonia ranks least vulnerable in this small-states context because of larger population, fertile land, proximity to many neighbors, and relatively high total GNI.

### VIII. Empirical link to real economic outcomes (regressions)
- Dependent variable: annualized nominal growth in GNI measured in US dollar terms between 2001 and 2010 (annual data).
- Sample notes:
  - Nine observations for most countries; Montenegro, Sao Tome and Principe, Timor-Leste have fewer; Gibraltar excluded due to lack of time series data.
  - Cross-correlation likely due to the recent financial crisis affecting all countries.
- Separate regressions fitted using each of the five vulnerability indicator rankings and the aggregate index. Exact slope coefficients, P-values and RMSEs:
  - population: Slope coefficient -0.19, P-value 0.0006 **, RMSE(ith) 5.15, RMSE(ith)/RMSE(index) 1.0158
  - land: Slope coefficient -0.16, P-value 0.0063 **, RMSE(ith) 5.40, RMSE(ith)/RMSE(index) 1.0651
  - distance: Slope coefficient -0.18, P-value 0.0015 **, RMSE(ith) 5.25, RMSE(ith)/RMSE(index) 1.0355
  - income: Slope coefficient -0.14, P-value 0.0167 *, RMSE(ith) 5.50, RMSE(ith)/RMSE(index) 1.0848
  - income/capita: Slope coefficient 0.02, P-value 0.6899, RMSE(ith) 5.84, RMSE(ith)/RMSE(index) 1.1519
  - index: Slope coefficient -0.32, P-value 0.0003 **, RMSE(ith) 5.07, RMSE(ith)/RMSE(index) 1.0000
- Memorandum item:
  - index-ex (index excluding income per capita and dropping outliers Timor-Leste and Equatorial Guinea): Slope coefficient -0.23, P-value 0.0000 **, RMSE(ith) 2.96, RMSE(ith)/RMSE(index) 0.5838
- Key empirical findings:
  - Population size, distance to the nearest continent, and fertility of land are the most statistically significant explanatory variables for growth (negative slope coefficients).
  - Income per capita shows a weak positive relationship with growth (slope 0.02) and is statistically insignificant (P-value 0.6899).
  - The aggregate index yields better fit than individual indicators; excluding income per capita and outliers (index-ex) dramatically improves fit (RMSE falls to 2.96).

### IX. Interpretations, conclusions and policy-relevant points
- Main determinants of vulnerability and their role in economic outcomes:
  - Population size: smaller populations limit economies of scale in provision of public goods and services.
  - Distance/isolation: extreme physical remoteness raises trade and transport costs and limits market access.
  - Arable land: limited fertile land increases dependence on imported foodstuffs.
  - Total GNI: small GNI constrains ability to reap economies of scale.
- Policy-relevant point: combination of multiple vulnerabilities ("a confluence of vulnerabilities") is important in determining growth outcomes; states exposed to a larger number of adverse factors find it more difficult to achieve consistently good economic performance.
- Specific challenges for Pacific small island states unlikely to be resolved soon:
  - Physical barriers in trade, commerce, and labor mobility due to distance.
  - Environmental hazards linked to geography: tsunamis, hurricanes, and rising sea levels.

### X. Appendix I — country classifications and selected indicators (selected entries)
- Columns defined:
  - Country name; Regional; Population (Thousands) — As @ 2010; Arable land (m2/capita) — As @ 2009; EVI-13 Index (1-7) — As @ 2010; Distance to closest continent (km); Closest continent; GNI, Current prices, US$ million — As @ 2010; GNI/Capita, Current prices, US$ — As @ 2010; GNI/Capita, PPP basis, US$ — As @ 2010
- Selected country rows (exact values as presented):
  - Tuvalu; East Asia & Pacific; 10; 0.0; 7; 3,300; Australia; 47; 4,760; 5,351
  - Palau; East Asia & Pacific; 20; 488.5; 6; 1,463; Asia; 134; 6,560; 11,000
  - Gibraltar; Europe & Central Asia; 29; 0.0; 1; 0; Europe; 969; 33,151; 38,217
  - Monaco; Europe & Central Asia; 35; 0.0; 1; 0; Europe; 6,479; 183,150; 182,990
  - Marshall Islands; East Asia & Pacific; 54; 370.1; 7; 3,500; Australia; 197; 3,640; 3,003
  - Samoa; East Asia & Pacific; 183; 1,365.5; 7; 3,800; Australia; 549; 3,000; 4,270
  - Fiji; East Asia & Pacific; 861; 1,859.1; 7; 2,600; Australia; 3,123; 3,630; 4,510
  - Macao SAR, China; East Asia & Pacific; 544; 0.0; 2; 2; Asia; 18,527; 34,880; 45,220
  - Luxembourg; Europe & Central Asia; 506; 1,225.7; 1; 0; Europe; 39,030; 77,160; 61,790
  - Guyana; Latin America & Caribbean; 754; 5,566.7; 1; 0; South America; 2,164; 2,870; 3,450
  - Trinidad and Tobago; Latin America & Caribbean; 1,341; 186.4; 2; 20; South America; 20,625; 15,380; 24,040

### XI. Data sources and metadata (definitions and provenance)
- Arable Land: Sourced from The World Bank, Open Data as at 2010. Definition: includes land under temporary crops (double-cropped areas are counted once), temporary meadows for mowing or for pasture, land under market or kitchen gardens, and land temporarily fallow. Land abandoned as a result of shifting cultivation is excluded.
- Environmental Vulnerability Index, EVI-13; Distance to Closest Continent; Closest Continent: Sourced from SOPAC and UNEP, 2005. Note: Distance captures proximity to the nearest continent; if a country is within a continent, this value is zero.
- Gross National Income (GNI), current US dollars; GNI per capita, current US dollars; GNI per capita, PPP: Sourced from The World Bank, Open Data as at 2010.
- Additional notes: Some values estimated by author using GDP and GDP per capita data from the CIA World Factbook. Population sourced from The World Bank, Open Data as at 2010. Regional classification of states sourced from The World Bank, as at April 2012. Land Area sourced from The World Bank, Open Data as at 2010.

*Source: _wp12223 - Appendix I. Country Classifications and Data; PDF chapter/section.*

### Appendix I. Country Classifications and Data ........................................................................23

### _wp12223 - Appendix I. Country Classifications and Data ........................................................................23

### Appendix I. Country Classifications and Data
- Appendix I. Country Classifications and Data ........................................................................23

### Appendix II. Data Sources and Metadata
- Appendix II. Data Sources and Metadata ................................................................................24

*Source: _wp12223 - Appendix I. Country Classifications and Data; PDF chapter/section.*

### References ................................................................................................

### References

### I. Introduction
- Purpose: consider features of small island states in the Asia-Pacific region and document key characteristics that set them apart from small states in other regions.
- Scope: restricted to macro-oriented indicators—population size, income per capita, fertility of land, ability to tap into economies of scale, and degree of geographic isolation.
- Exclusions: micro-oriented variables (telecommunications, electricity generation) and development indicators (literacy, infant mortality) are left aside.
- Core finding preview: Pacific small island states are typically extremely geographically isolated, have limited scope to tap into economies of scale due to small populations, face complications interpreting income per capita, and often lack fertile land constraining import substitution through agriculture.
- Organization: Sections II–VI discuss indicators; Section VII summarizes via ranking and relates to economic outcomes; Section VIII concludes.

### II. Defining small states by population
- Common metric: population often used to define smallness; World Bank defines a small state as one with a population of less than 1½ million people (World Bank, 2007).
- Sample used: 50 small states (World Bank data).
  - Examples: Tuvalu with a population of only around 10,000 people; Trinidad and Tobago with around 1.3 million people.
- Note on heterogeneity:
  - Microstates can be heterogeneous across economic indicators; population alone may not predict income.
  - Example: Monaco’s population is around three times Tuvalu’s, but GNI per capita is almost 40 times larger.
- Data exclusions: 15 territories excluded from initial World Bank sample of 65 due to unavailable comparable data (list provided in source).

### III. Lacking economies of scale
- General relationship:
  - Population positively related to total income (labor input), but trend has weak explanatory power (Figure 2, R² = 0.244).
  - Outliers exist (e.g., Luxembourg).
- Constraints for Pacific small islands:
  - Lack attributes of financial centers or trade hubs (proximity to large skilled labor markets, infrastructure sharing, strategic shipping routes).
  - Geography precludes sharing infrastructure; duplication required for dispersed populations.
  - Some financial-service activities less tied to location (business registrations, ship registration) remain feasible.
- A. Fixed costs spread over narrow base
  - Small populations imply small total income and narrow tax base; fixed costs for public goods/services can be an unusually high share of national income.
  - Consequences: insufficient tax revenue for basic health services, transportation, government administration; small administrations with limited capacity.
  - Disaster-resilient infrastructure increases fixed-cost burden.
- B. Scope for regional cooperation
  - Regional aggregation of populations could create larger common markets or shared infrastructure.
  - Aggregate-population illustration: Pacific islands’ regional aggregate population is not clearly disadvantaged relative to other regions (Figure 3).
  - Barriers to cooperation: cultural, language, distance, legal, political heterogeneity.
  - Existing regional mechanisms: Pacific Islands Forum, Forum Economic Ministers Meeting (FEMM), Pacific Plan, Pacific Agreement on Closer Economic Relations, Pacific Financial Technical Assistance Center (PFTAC).

- Key statistics/examples (percentage deviation from small-state median GNI; median GNI of small state sample was US$1,719 million, 2010):
  - Tuvalu: Per cent below median = -97.3
  - Palau: Per cent below median = -92.2
  - Monaco: Per cent above median = 276.9
  - Luxembourg: Per cent above median = 2,170.6
  - Channel Islands: Per cent above median = 495.8
  - Note: Of Pacific small countries, only Timor-Leste, Macau, and Fiji had GNI larger than the median US$1.7 billion in 2010.

### IV. Income per capita
- Importance: commonly used as proxy for poverty and aid eligibility (IDA eligibility, IMF PRGT linkage).
- Limitations for microstates:
  - Extremely small populations can inflate income per capita if numerator (total income) shared across very small denominator; may mask inadequate total income to fund domestic services.
  - Example mechanism: high per-capita income but expensive out-of-country medical services reduce disposable per-capita consumption and welfare.
  - Increasing distortions as denominator approaches extremely small values.
- Distributional note:
  - Top decile of small states by GNI per capita is entirely European (Monaco, Liechtenstein, Luxembourg, Channel Islands, San Marino).
  - Proximity to neighboring labor markets and infrastructure helps some microstates overcome smallness.
- Policy implication: multiple vulnerabilities (small population, geographic isolation, low land fertility) compound, reducing the ability to overcome disadvantages.

### V. Scope for import substitution through agriculture
- Common feature: many small states have limited fertile/arable land despite sometimes large total landmass.
- Risk channel: limited arable land → heavy dependence on food imports → greater exposure to balance of payments crises when food prices rise.
- Examples:
  - Five states in sample have no arable land at all: Macao, Greenland, Monaco, Gibraltar, Tuvalu.
  - Tuvalu: Arable land m2 per capita = 0.
  - Extremes in arable land per capita (Table 2 highlights):
    - Guyana: Arable Land m2 per capita = 5,567
    - Bahrain: Arable Land m2 per capita = 8
  - Some Pacific islands (Fiji, Samoa, Tonga, Timor-Leste) are relatively fertile compared with many other small states.
- Regional patterns:
  - Pacific states generally more infertile than small countries in Europe or Latin America and the Caribbean, but not as infertile as many Middle East and North Africa states or Greenland.
  - Arable land as percent of total landmass varies widely (Figure 6 cited).
- Trade as mitigation:
  - Trade with agriculturally productive partners could offset domestic constraints, but geographic isolation raises transport costs and limits market access.
  - Fishing rights and fisheries exports provide possible revenue sources for Pacific islands, but isolation remains a constraint.

### VI. Geographic isolation
- Defining feature of small Pacific islands: extreme remoteness from continents and often from neighboring countries.
- Implications:
  - Higher transport costs for trade and commerce as distance increases.
  - Increased vulnerability from import dependence (especially food) and the costliness of accessing external services and markets.
  - Tourism viability is conditional on specific comparative advantages; distance and cost disadvantages limit feasibility.
  - Dispersion across islands increases susceptibility to natural disasters (tsunamis, hurricanes) and complicates resilient infrastructure provision.
  - Environmental vulnerability: rising sea levels and global warming disproportionately affect some Pacific islands; similar environmental pressures affect Caribbean small states.

*Source: References section and main text of the supplied IMF working paper content.*

### Appendix I lists some relevant indicators such as the isolation sub

### _wp12223 - Appendix I lists some relevant indicators such as the isolation sub

### Isolation and geographic vulnerability
- Appendix I references the isolation sub index (EVI-13) of the Environmental Vulnerability Index calculated by SOPAC and the UNEP.
- In the sample of small states, Samoa, Marshall Islands, Tuvalu, Kiribati, and Tonga are among the most isolated states in the world.
- Each of these Pacific states is "more than 3,000 kilometers from the nearest continent, Australia."
- Isolation raises trade costs and vulnerability because:
  - Many remote countries do not export goods, leaving container ships without rolling stock for return legs, increasing delivery costs.
  - Smaller island countries often lack infrastructure for larger container ships to dock.
  - Long maritime distances amplify exposure to fossil fuel price fluctuations (fuel is a major part of shipping costs).

### Distance metrics and regional comparisons
- Pacific islands’ average distance to the nearest continent is "more than four to five times" that of the average country in the Caribbean or Sub-Saharan Africa.
- Small states in Europe or Northern Africa and the Middle East are considerably less isolated by the measure used.
- Mayer and Zignago (2011) provide a more comprehensive remoteness measure (distance to all other countries in a 224-country sample), which further supports that small Pacific island states are particularly isolated because they are dispersed and far from each other and from large markets.
- Augmenting distance metrics by weighting distances by GDP (Chen et al, 2012) makes Pacific small states appear even worse off because large landmasses like Australia represent relatively small markets in per-capita terms for some small states.

### Vulnerability ranking methodology
- Five indicators used to rank 50 small states: population, arable land per capita, distance to nearest continent, total GNI (US dollar level), and GNI per capita.
- Ranking convention: higher rank number indicates greater relative vulnerability on that indicator (e.g., Tuvalu rank 50 on population = most vulnerable by population).
- An aggregate summary ranking (the index) is the equally weighted average across the five individual indicator rankings.
- Caveats noted:
  - Mechanical indices cannot fully reflect complexities and interactions among variables.
  - Weighting equally implies no reasoned preference function; alternative weightings could alter outcomes.

### Table 3 — Selected ranking outcomes (rank averages)
- Top vulnerability by index average:
  - Tuvalu: Rank average 46.2 (Population 50, Arable land 50, Distance 48, GNI 50, GNI/Capita 33)
  - Kiribati: Rank average 41.6 (36, 34, 47, 46, 45)
  - Marshall Islands: Rank average 40.6 (43, 28, 49, 48, 35)
  - Micronesia, Fed. Sts.: Rank average 40.0 (32, 38, 44, 45, 41)
  - Palau: Rank average 38.4 (49, 26, 39, 49, 29)
- Lower vulnerability by index average:
  - Luxembourg: Rank average 8.0 (18, 11, 7, 1, 3)
  - Estonia: Rank average 6.0 (2, 2, 2, 5, 19)
  - Cyprus: Rank average 12.4 (7, 20, 21, 2, 12)
  - Montenegro: Rank average 12.6 (14, 3, 3, 16, 27)
  - Guyana and Trinidad and Tobago: Rank average 15.2 each (Guyana: 10, 1, 1, 24, 40; Trinidad and Tobago: 1, 37, 18, 3, 17)
- Author’s interpretation:
  - Tuvalu is the most vulnerable small state in the sample due to very small population, no arable land, high isolation, and small GNI, outweighing its relatively favorable income per capita ranking.
  - Estonia ranks least vulnerable in this small-states context because of larger population, fertile land, proximity to many neighbors, and relatively high total GNI.

### Regional summary observations
- Small States in the Pacific:
  - Are "the most vulnerable on a number of counts considered in this paper."
  - Common vulnerabilities include isolation, inability to reap economies of scale, generally low income per capita, small populations, and little arable land.
  - They rank worse than the sample average of "25.5" on all measures considered.
- Small states in Latin America and the Caribbean, South Asia, and Sub-Saharan Africa are "probably somewhat less vulnerable" on the factors considered, showing a common degree of overall exposure.
- Small countries in Europe appear least disadvantaged in this sample, influenced by outliers that are highly developed and rich despite small populations.

### Empirical link to real economic outcomes (regressions)
- Dependent variable: annualized nominal growth in GNI measured in US dollar terms between 2001 and 2010 (annual data).
- Sample issues:
  - Nine observations for most countries; Montenegro, Sao Tome and Principe, Timor-Leste have fewer; Gibraltar excluded due to lack of time series data.
  - High likelihood of cross-correlation due to the recent financial crisis affecting all countries.
- Separate regressions fit using each of the five vulnerability indicator rankings and the aggregate index.
- Table 4 — Slope coefficients, P-values, RMSEs (exact values):
  - population: Slope coefficient -0.19, P-value 0.0006 **, RMSE(ith) 5.15, RMSE(ith)/RMSE(index) 1.0158
  - land: Slope coefficient -0.16, P-value 0.0063 **, RMSE(ith) 5.40, RMSE(ith)/RMSE(index) 1.0651
  - distance: Slope coefficient -0.18, P-value 0.0015 **, RMSE(ith) 5.25, RMSE(ith)/RMSE(index) 1.0355
  - income: Slope coefficient -0.14, P-value 0.0167 *, RMSE(ith) 5.50, RMSE(ith)/RMSE(index) 1.0848
  - income/capita: Slope coefficient 0.02, P-value 0.6899, RMSE(ith) 5.84, RMSE(ith)/RMSE(index) 1.1519
  - index: Slope coefficient -0.32, P-value 0.0003 **, RMSE(ith) 5.07, RMSE(ith)/RMSE(index) 1.0000
- Memorandum item (index-ex: index excluding income per capita and dropping outliers Timor-Leste and Equatorial Guinea):
  - index-ex: Slope coefficient -0.23, P-value 0.0000 **, RMSE(ith) 2.96, RMSE(ith)/RMSE(index) 0.5838
- Key empirical findings:
  - The slope coefficients for population, distance, land, and income are negative (consistent with expectation that higher relative vulnerability associates with lower growth).
  - Income per capita shows a weak positive relationship with growth (slope 0.02) and is statistically insignificant (P-value 0.6899).
  - The most statistically significant explanatory variables (P-value) for growth are population size, distance to the nearest continent, and fertility of land.
  - The aggregate index yields better fit than individual indicators; excluding income per capita and outliers (index-ex) dramatically improves fit (RMSE falls to 2.96).

### Interpretations and conclusions
- The fitted relationships generally indicate the expected negative relationship between degree of vulnerability and growth, though results are affected by outliers (Timor-Leste, Equatorial Guinea).
- Clustering of Pacific states in plots suggests their high vulnerability may impede economic performance by dragging average growth lower.
- Main determinants of vulnerability and their role in economic outcomes:
  - Population size: smaller populations limit economies of scale in provision of public goods and services.
  - Distance/isolation: extreme physical remoteness raises trade and transport costs and limits market access.
  - Arable land: limited fertile land increases dependence on imported foodstuffs.
  - Total GNI: small GNI constrains ability to reap economies of scale.
- Policy-relevant point: combination of multiple vulnerabilities ("a confluence of vulnerabilities") is important in determining growth outcomes; states exposed to a larger number of adverse factors find it more difficult to achieve consistently good economic performance.
- Specific challenges for Pacific small island states that are unlikely to be resolved soon:
  - Physical barriers in trade, commerce, and labor mobility due to distance.
  - Environmental hazards linked to geography: tsunamis, hurricanes, and rising sea levels.

*Source: SOPAC; Author’s calculations (text provided).*

### Appendix I. Country Classifications and Data

### Appendix I. Country Classifications and Data

### Small and Vulnerable States — country classifications and indicators
- Columns and definitions (as presented):
  - Country name
  - Regional
  - Population (Thousands) — As @ 2010
  - Arable land (m2/capita) — As @ 2009
  - EVI-13 Index (1-7) — As @ 2010
  - Distance to closest continent (km) — Isolation to closest continent
  - Closest continent — Name
  - GNI, Current prices, US$ million — As @ 2010
  - GNI/Capita, Current prices, US$ — As @ 2010
  - GNI/Capita, PPP basis, US$ — As @ 2010

- Country entries (as presented):
  - Tuvalu; East Asia & Pacific; 10; 0.0; 7; 3,300; Australia; 47; 4,760; 5,351
  - Palau; East Asia & Pacific; 20; 488.5; 6; 1,463; Asia; 134; 6,560; 11,000
  - Gibraltar; Europe & Central Asia; 29; 0.0; 1; 0; Europe; 969; 33,151; 38,217
  - San Marino; Europe & Central Asia; 32; 317.1; 1; 0; Europe; 1,572; 49,866; 35,163
  - Monaco; Europe & Central Asia; 35; 0.0; 1; 0; Europe; 6,479; 183,150; 182,990
  - Liechtenstein; Europe & Central Asia; 36; 832.6; 1; 0; Europe; 4,903; 137,070; 151,189
  - St. Kitts and Nevis; Latin America & Caribbean; 52; 763.3; 7; 2,100; South America; 615; 11,740; 15,850
  - Marshall Islands; East Asia & Pacific; 54; 370.1; 7; 3,500; Australia; 197; 3,640; 3,003
  - Greenland; Greenland; 57; 0.0; 5; 750; North America; 1,466; 26,020; 25,604
  - Dominica; Latin America & Caribbean; 68; 885.5; 5; 697.5; South America; 458; 6,760; 11,990
  - Isle of Man; Europe & Central Asia; 83; 603.4; 2; 30; Europe; 3,972; 48,910; 47,926
  - Andorra; Europe & Central Asia; 85; 117.8; 1; 0; Europe; 3,447; 41,750; 36,113
  - Seychelles; Sub-Saharan Africa; 87; 115.6; 5; 475; Africa; 845; 9,760; 21,210
  - Antigua and Barbuda; Latin America & Caribbean; 89; 901.8; 7; 2,110; South America; 1,169; 13,170; 20,240
  - Kiribati; East Asia & Pacific; 100; 200.9; 7; 3,250; Australia; 200; 2,010; 3,530
  - Tonga; East Asia & Pacific; 104; 1,537.6; 7; 3,200; Australia; 342; 3,280; 4,580
  - Grenada; Latin America & Caribbean; 104; 191.4; 4; 140; South America; 724; 6,930; 9,890
  - St. Vincent and the Grenadines; Latin America & Caribbean; 109; 457.3; 4; 300; South America; 688; 6,300; 10,830
  - Micronesia, Fed. Sts.; East Asia & Pacific; 111; 180.1; 7; 2,500; Australia; 304; 2,730; 3,490
  - Channel Islands; Europe & Central Asia; 153; 260.8; 2; 30; Europe; 10,241; 67,960; 66,781
  - Sao Tome and Principe; Sub-Saharan Africa; 165; 604.6; 4; 225; Africa; 199; 1,200; 1,920
  - St. Lucia; Latin America & Caribbean; 174; 172.4; 5; 697.5; South America; 1,142; 6,560; 10,520
  - Samoa; East Asia & Pacific; 183; 1,365.5; 7; 3,800; Australia; 549; 3,000; 4,270
  - Vanuatu; East Asia & Pacific; 240; 834.5; 7; 1,800; Australia; 633; 2,640; 4,320
  - Barbados; Latin America & Caribbean; 273; 585.4; 4; 350; South America; 3,454; 12,660; 19,000
  - Maldives; South Asia; 316; 126.6; 5; 500; Asia; 1,818; 5,750; 8,110
  - Iceland; Europe & Central Asia; 317; 220.5; 6; 1,000; Europe; 10,381; 32,710; 27,680
  - Bahamas, The; Latin America & Caribbean; 343; 233.3; 4; 200; North America; 6,973; 20,610; 24,800
  - Belize; Latin America & Caribbean; 345; 2,030.8; 1; 0; North America; 1,313; 3,810; 6,210
  - Brunei Darussalam; East Asia & Pacific; 399; 75.2; 6; 950; Asia; 12,461; 31,800; 50,180
  - Malta; Europe & Central Asia; 413; 193.7; 4; 250; Europe; 7,958; 19,270; 24,840
  - Cape Verde; Sub-Saharan Africa; 496; 1,209.7; 5; 580; Africa; 1,620; 3,270; 3,820
  - Luxembourg; Europe & Central Asia; 506; 1,225.7; 1; 0; Europe; 39,030; 77,160; 61,790
  - Suriname; Latin America & Caribbean; 525; 1,105.5; 1; 0; South America; 3,077; 5,920; 7,680
  - Solomon Islands; East Asia & Pacific; 538; 297.3; 7; 1,950; Australia; 552; 1,030; 2,210
  - Macao SAR, China; East Asia & Pacific; 544; 0.0; 2; 2; Asia; 18,527; 34,880; 45,220
  - Montenegro; Europe & Central Asia; 631; 2,739.6; 1; 0; Europe; 4,260; 6,750; 12,930
  - Equatorial Guinea; Sub-Saharan Africa; 700; 1,884.6; 1; 0; Africa; 10,182; 14,540; 23,750
  - Bhutan; South Asia; 726; 1,033.1; 1; 0; Asia; 1,361; 1,870; 4,990
  - Comoros; Sub-Saharan Africa; 735; 1,088.8; 4; 300; Africa; 550; 750; 1,090
  - Guyana; Latin America & Caribbean; 754; 5,566.7; 1; 0; South America; 2,164; 2,870; 3,450
  - Fiji; East Asia & Pacific; 861; 1,859.1; 7; 2,600; Australia; 3,123; 3,630; 4,510
  - Djibouti; Middle East & North Africa; 889; 22.5; 1; 0; Africa; 1,105; 1,270; 2,460
  - Cyprus; Europe & Central Asia; 1,104; 788.3; 3; 60; Asia; 23,655; 29,430; 30,300
  - Timor-Leste; East Asia & Pacific; 1,124; 1,467.5; 4; 600; Australia; 2,493; 2,220; 3,600
  - Swaziland; Sub-Saharan Africa; 1,186; 1,475.5; 1; 0; Africa; 3,119; 2,630; 4,840
  - Bahrain; Middle East & North Africa; 1,262; 7.9; 2; 15; Asia; 19,714; 18,730; 24,710
  - Mauritius; Sub-Saharan Africa; 1,281; 679.0; 6; 1,400; Africa; 9,925; 7,750; 13,960
  - Estonia; Europe & Central Asia; 1,340; 4,448.9; 1; 0; Europe; 19,371; 14,460; 19,760
  - Trinidad and Tobago; Latin America & Caribbean; 1,341; 186.4; 2; 20; South America; 20,625; 15,380; 24,040

### Data sources and metadata — definitions and provenance
- Arable Land:
  - Sourced from The World Bank, Open Data as at 2010. Available via Internet: http://data.worldbank.org/
  - Definition: includes land under temporary crops (double-cropped areas are counted once), temporary meadows for mowing or for pasture, land under market or kitchen gardens, and land temporarily fallow. Land abandoned as a result of shifting cultivation is excluded.
- Environmental Vulnerability Index, EVI-13; Distance to Closest Continent; Closest Continent:
  - Sourced from SOPAC and UNEP, 2005.
  - Note: Distance captures proximity to the nearest continent; if a country is within a continent, this value is zero. Isolated countries may have greater risk of loss of ecosystem types and species during periods of stress, support fewer species, and have less chance of genetic interchange.
- Gross National Income (GNI), current US dollars:
  - Sourced from The World Bank, Open Data as at 2010. Available via Internet: http://data.worldbank.org/
  - Definition: Sum of value added by all resident producers plus any product taxes (less subsidies) not included in the valuation of output plus net receipts of primary income from abroad. Data in current U.S. dollars. World Bank Atlas method noted for conversion and smoothing of exchange rate and price fluctuations.
- Gross National Income (GNI) per capita, current US dollars:
  - Sourced from The World Bank, Open Data as at 2010. Available via Internet: http://data.worldbank.org/
  - Definition: GNI converted to U.S. dollars using the World Bank Atlas method, divided by the midyear population.
- Gross National Income (GNI) per capita, Purchasing Power Parity (PPP):
  - Sourced from The World Bank, Open Data as at 2010. Available via Internet: http://data.worldbank.org/
  - Definition: GNI converted to international dollars using purchasing power parity rates; data in current international dollars.
- Additional notes:
  - Some values estimated by author using GDP and GDP per capita data from the CIA World Factbook. Available via Internet: https://www.cia.gov/library/publications/the-world-factbook/
  - Population sourced from The World Bank, Open Data as at 2010 (midyear estimates, de facto definition).
  - Regional classification of states sourced from The World Bank, as at April 2012.
  - Land Area sourced from The World Bank, Open Data as at 2010.

*Source: Appendix I. Country Classifications and Data (as provided in the supplied content).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12223.pdf_
