## _wp13104

## Source details

**Canonical URL:** [_wp13104](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13104.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13104.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13104.pdf.json)

---

### I. Introduction
- Economic recovery in most Pacific island countries (PICs) following the global financial crisis has been weak.
- Two resource-rich countries, Papua New Guinea (PNG) and Solomon Islands, have rebounded strongly on the back of high commodity prices.
- Most other countries: annual GDP growth averaged just 0.3 percent during the period 2008-2010.
- At least six PICs out of the eleven IMF members have yet to reach their pre-crisis per capita GDP levels.
- Core research questions:
  - To what extent has economic growth in PICs been slower than other small states and other similar country groups?
  - Have natural conditions (small size, geographical location) and external shocks (fuel prices, changing trade preferences) played a role in PICs’ long-run growth?
  - Why has growth in PICs slowed over the past decade?
  - How might economic and other policies have affected PICs’ growth performance?
- Scope: unique dataset for small states; long-run perspective; case studies; comparison with other small states and country groups.

### II. Growth performance — stylized facts
- Long-run patterns:
  - 1970s: PIC economies grew at just below 4 percent per year.
  - 1990s: PIC growth rebounded and nearly caught up with LICs.
  - 2000s: sharp divergence; PICs had the weakest growth in four decades, managing to grow at only one-third the rate of LICs.
- Cross-country variation:
  - PNG and Solomon Islands: dramatic improvement in the last decade due to commodity booms.
  - Fiji: after historically outperforming neighbors, registered one of the slowest growth rates in the past decade largely due to domestic developments.
- Volatility:
  - Growth in PICs has been volatile but on average exhibited similar levels of volatility as other developing-country groups.
  - 2000s: PICs had lower growth volatility (along with the slowest growth rate) compared with earlier decades.

### III. Model, data and methodologies
- Model specification and calibration:
  - Growth-initial level (beta-convergence) model; core equation expressed in equations (1) and (2).
  - Calibration: set g + δ = 5% (adopted as g + d = 0.05 in methodology notes).
- Data coverage and units:
  - Cross-section dataset covering period 1992-2008 for 45 small states (some regressions cover 40 or 39 due to data limitations).
  - All money values in constant 2005 prices in US dollars.
  - Dependent variable gy = average real growth of per capita GDP over 1992-2008.
- Main explanatory variables (averaged over 1992-2008 unless noted):
  - ln(y0): GDP per capita in 1992.
  - ln(investment rate): investment-to-GDP (%) averaged over 1992-2008.
  - ln(n + g + d): population growth plus g + δ (%) averaged over 1992-2008.
  - Aid/GNI: foreign aid-to-GNI ratio (%) — most recent data (2007) used.
  - Political stability: World Bank Worldwide Governance Indicators (averaged over 1996-2008).
  - Exports/GDP: exports-to-GDP ratio (%) averaged over 1992-2008.
  - Growth volatility: standard deviation of GDP growth rates over the period.
  - ln(distance): GDP-weighted distance to capital cities of major overseas markets.
  - Regional and special dummies (including Equatorial Guinea).
- Estimation and identification:
  - Regressions 1–3 use ordinary least squares (OLS); Regression 4 uses two-stage least squares (2SLS).
  - Growth volatility found endogenous; political stability and exports-to-GDP used as instruments.
  - Instrument validity: Sargan statistic reported as 0.7387 (regression 4).
  - Endogeneity tests show no evidence of endogeneity for investment and aid (Sargan, Hausman tests).
  - Collinearity diagnostics: mean VIF 2.33 (regressions 1 and 2), 1.99 (regression 3).
  - Heteroskedasticity tests: Breusch-Pagan p-values 0.3634 (regression 1), 0.4534 (regression 2), 0.7104 (regression 3).
  - All regressors significant at least at the 10% level, with most significant at 1% or 5% in the reported specifications.

### IV. Determinants of growth in small states — key empirical findings
- Geography and remoteness:
  - Geography has a large influence: PICs suffer a distance-related disadvantage in per capita GDP growth of about 1½ percentage points compared with an average non-Pacific small state.
  - PICs’ annual average per capita GDP growth over 1992-2008 was a little over 0.7 percent; without geographical disadvantage, PICs could have grown more than three times as fast during the period.
- Multivariate regression (selected coefficients and t-stats reproduced exactly as in source, based on regressions 2 and 3 in Table 1):
  - Pacific constant: -3.13  [-3.33]
  - ln(y0): 7.00  [1.56]; regression 2: -1.90  [-2.11]; regression 3: -1.36  [2.90]
  - ln(invest. rate): -0.13  [-0.58]; regression 2: 15.68  [3.47]; regression 3: 1.08  [-4.77]
  - ln(n + g + d): 1.23  [2.11]; regression 2: -1.30  [-3.80]; regression 3: -3.94  [2.20]
  - Aid/GNI: -3.21  [-2.04]; regression 2: 1.05  [1.78]; regression 3: -0.06  [-3.32]
  - Political stability: -1.90  [-2.11]; regression 2: -0.05  [-2.61]; regression 3: -0.07  [-2.44]
  - Exports/GDP: 15.68  [3.47]; regression 2: 1.25  [2.50]; regression 3: 1.41  [3.19]
  - Growth volatility: -1.30  [-3.80]; regression 2: 0.03  [2.07]; regression 3: 0.04  [2.79]
  - ln(distance): 1.05  [2.11]; regression 2: 35.09  [2.90]; regression 3: -2.31  [-1.91]
- Sample sizes and fit statistics:
  - Sample size: 45 (regression 1), 40 (regression 2), 39 (regression 3), 40 (2SLS regression 4).
  - Adjusted/centered R2: 0.8161 (regression 1), 0.8786 (regression 2), 0.9068 (regression 3), 0.8818 (regression 4).
- Interpretation and sensitivity:
  - Distance (remoteness) is a principal source of PICs’ growth disadvantage among small states.
  - Geography alone does not fully explain slower PIC growth; other determinants include exports openness, political stability, volatility, investment, and aid, with signs and significance changing across model specifications.
  - The impact and statistical significance of variables are sensitive to model choice and controls.

### V. Main determinants of PIC per capita GDP growth (relative to other small states) — quantitative magnitudes
- Convergence (initial income):
  - PICs were on average poorer in 1992; this lower initial income allowed PICs to grow about ½ percentage point faster than an average non-Pacific small state.
- Investment:
  - Over 1992-2008, investment in PICs averaged 22½ percent of GDP, about 6 percentage points lower than the average of all small states.
  - Had PICs achieved the average investment rate of non-Pacific small states, real per capita GDP growth would have been about ¼ percentage point higher.
  - Example: Fiji’s investment rate is 14 percent of GDP while that of Vanuatu is 20 percent; this translates to a difference in the growth rate of ¼ percentage point.
- Exports-to-GDP (export openness):
  - Each 10 percentage point increase in the exports-to-GDP ratio raises per capita GDP growth by about 0.3 percentage point.
  - The exports-to-GDP ratio in PICs is 24 percentage points lower than the average of non-Pacific small states, implying PICs could have grown by 0.6 percentage point faster had they exported as much as other small states (in percent of GDP).
  - Greater export openness increases output volatility: each 10 percentage point change in export openness is associated with 0.2 percentage point change in the standard deviation of GDP growth.
  - Less than a quarter of the growth benefits from greater export openness is nullified by the associated increase in output volatility, leaving PICs worse off by about 0.5 percentage point from their lower openness compared with other small states.
- Imports and trade turnover:
  - Greater imports as percent of GDP have no statistically significant impact on growth.
  - Openness measured by trade turnover (exports plus imports) to GDP shows no positive impact on growth in reported estimations.
- Aid:
  - For each 10 percentage point increase in aid as percent of GDP, growth is lower by 0.6 percentage point.
  - Preliminary tests indicate no reverse causality (slower growth does not lead to more aid); the relationship applies to total aid (no disaggregation by type available).
  - Negative relationship should not be interpreted as aid lowering welfare; much aid is aimed at reducing poverty rather than increasing economic growth.
- Political stability:
  - PICs score more favorably on the World Bank Political Stability index than other small states, giving PICs, on average, an advantage of 0.3 percentage point in growth over other small states.
  - Political stability varies substantially among PICs; the highest-scoring country has a growth advantage of more than one percentage point over the lowest-scoring country.
- Growth volatility:
  - Measured by standard deviation of per capita GDP growth rate.
  - Political stability and exports-to-GDP affect growth both directly and indirectly via growth volatility.
- Decomposition summary:
  - Combined policy- and geography-related factors, together with other determinants, collectively lower per capita GDP growth in PICs by about another percentage point compared with an average small state.
  - Specific magnitude highlights:
    - Initial income: +½ percentage point advantage for PICs (convergence effect).
    - Investment shortfall: −¼ percentage point from lower investment.
    - Exports shortfall: −0.6 percentage point from being 24 percentage points lower in exports-to-GDP.
    - Export-induced volatility: each 10 percentage point increase in export openness raises standard deviation by 0.2 percentage point.
    - Aid effect: −0.6 percentage point for each 10 percentage point increase in aid/GDP.
    - Political stability advantage: +0.3 percentage point on average across PICs.

### VI. Explaining the growth slowdown in the 2000s (2001-08)
- Period decomposition:
  - Sample split into 1992-2000 and 2001-08 using same regression coefficients to predict sub-period growth.
- Main contributors:
  - A decline in the exports-to-GDP ratio is a major contributor to the growth slowdown in the 2000s.
  - The average export openness ratio fell by as much as 4½ percentage points in the 2000s from its 1990s level.
  - Investment increased slightly and output volatility declined, which together should have added over half percentage point to growth, but the decline in export openness offset part of these gains.
- Competitiveness and REER:
  - Real exchange rate appreciation may have weakened competitiveness in several countries.
  - In the 2000s the REER appreciated for all six countries with their own currencies except Fiji.

### VII. Box: Samoa — Managing Shocks and Regaining Growth Momentum (case study)
- Background:
  - Samoa experienced strong growth until the mid 2000s; reforms in the mid-1990s and sound macro policies delivered an annual average GDP growth rate of 5 percent.
  - Investment related to the 2007 Pacific Games provided an additional boost.
- Shocks and outcomes:
  - Sequence of shocks: food and fuel price hikes; a devastating tsunami (2009); the global financial crisis (2008).
  - Despite a large boost to government expenditure in FY2009/10 - 2011/2012, economic growth slowed significantly.
  - Average overall fiscal deficit during this period was 7.3 percent of GDP.
  - Real GDP growth between 2006/07 to 2010/11 averaged 0.3 percent, the slowest since the mid 1960s.
  - Public debt reached well over 50 percent of GDP at the end of FY 2010/11.
- Sectoral notes:
  - Tourism and remittances have recovered only slowly.
  - Yazaki automotive component plant declining since 2008.
  - Agriculture has been stagnating in recent years.
  - Increase in the trade deficit largely due to imports for post-tsunami reconstruction; the strong Tala may have also boosted imports.
- Policy challenges and recommendations:
  - Need to wind down fiscal deficits while private sector has not yet taken up the slack.
  - Government committed to bringing down public debt over the medium term to maintain macroeconomic stability.
  - Fiscal consolidation is essential but may add headwinds to weak recovery.
  - Substantial efforts needed to strengthen competitiveness and revitalize private sector-led growth.
  - Structural reform should progress further to provide a more favorable environment for private-sector-led development.
  - Samoa’s strong reform track record and government commitment bode well for future progress.
- Key statistics (verbatim):
  - Annual average GDP growth rate in mid-1990s period: 5 percent.
  - Government expenditure boost period: FY2009/10 - 2011/2012.
  - Average overall fiscal deficit: 7.3 percent of GDP.
  - Real GDP growth, 2006/07 to 2010/11: averaged 0.3 percent.
  - Slowest growth since: the mid 1960s.
  - Public debt at end of FY 2010/11: well over 50 percent of GDP.
  - Notable shock events: food and fuel price hikes; a devastating tsunami (2009); the global financial crisis (2008).
  - Declining firm: Yazaki automotive component plant (declining since 2008).

### VIII. Transportation developments, trade costs and policy implications
- Transportation and technology:
  - Introduction of jet engines in the late 1950s and containerized shipping in the 1960s increased air and ocean transport efficiency.
  - Despite advances, ocean transport costs (as percent of the values of shipped goods) in the early 2000s were reported to be as they were in the 1950s (Hummels, 2007).
  - Rapid increases in transportation cost over the past decade, especially for air transportation, led to a sharp decline in the proportion of airlifted goods in trade.
  - For PICs, higher air transport costs have added substantially to transport costs of perishable products (fish, vegetables and fruits).
  - Containerized shipping was introduced in PICs mostly in the late 1970s and 1980s, later than elsewhere.
  - Limited competition in air and ocean transport and telecommunications raises costs to consumers (e.g., lack of competition in telecommunications raised costs in Samoa).
- Empirical findings:
  - A 10 percent increase in tourist arrivals per capita raises economic growth by about 0.2 percent (Thacker, 2011).
  - Weight-based measures overstate transport cost decline because shipped goods composition shifted to lighter, higher-valued goods.
  - Trade partners with 10 percent more exports enjoy 0.8 percent lower transport prices (Kleinert and Spies, 2011).
- Macroeconomic management recommendations:
  - Remoteness imposes a lower speed limit on PICs' growth and contributes to higher volatility; policymakers should set realistic growth targets and maintain larger policy buffers in good times.
  - Aid inflows can raise living standards via public services and cheaper imports but place pressure on export competitiveness through real exchange rate appreciation.
  - Recommendations to alleviate adverse effects of aid-supported strong exchange rates:
    - Direct more aid to productivity improvements.
    - Accelerate structural reforms.
  - Exchange rate regimes trade-offs:
    - Fixed pegs provide nominal anchors but can become unsustainable if inflation runs persistently higher than trading partners.
    - Flexible regimes can lead to rapid appreciation with large inflows; sterilization costly and often incomplete.
    - Foreign exchange markets are shallow; floating regimes may generate large exchange rate volatility harmful for trade.
  - Macroeconomic coordination:
    - Under fixed rates, monetary policy must be subordinate to preserve competitiveness; fiscal prudence is necessary.
    - Large windfall inflows could be saved (e.g., through a sovereign wealth fund) for inter-generational distribution and consumption smoothing.
- Structural reforms and integration:
  - Raising productivity faster than trading partners is essential under persistent inflation differentials.
  - Productivity growth in PICs has been generally slow; low investment is a key constraint.
    - Example: Total factor productivity growth in Fiji over 1983–2007 estimated at 0.5 percent per year (Chen and Singh, 2011).
  - Reducing cost of distance requires scaling up investment in transport and communication infrastructure and ensuring rigorous competition.
  - Greater regional cooperation (e.g., Pacific Forum Line; Pacific Islands Telecommunications Association) can help mitigate diseconomies of scale.
  - Further integration with metropolitan countries and Asia (including labor service trade and PACER-Plus) can provide larger markets and investment opportunities.

### IX. Fiji — lessons and hopes (selected findings)
- Key statistics:
  - Fiji averaged real GDP growth of 2.1 percent over the past two decades.
  - Since 1987, Fiji experienced four coups; coups had an immediate negative effect on the economy.
  - Private investment has been declining and volatile since 1988.
  - Public debt reached 51.5 percent of GDP.
  - A 2009 devaluation boosted the export sector, especially tourism, and helped restore macroeconomic stability.
  - The economy grew 2 percent in 2011.
  - Total factor productivity growth in Fiji over 1983–2007 estimated at 0.5 percent per year.
- Structural issues and reforms:
  - Declining public investment and maintenance of infrastructure until recently.
  - High and persistent inflation relative to trading partners under a pegged exchange rate eroded competitiveness.
  - Uncertainty over renewal of land leases reduced investment and productivity in land-based industries.
  - Mismanagement of the Fiji Sugar Corporation accelerated decline of the sugar industry and strained public finances.
  - Reforms undertaken include creation of a land bank, relaxation of foreign exchange controls, and reduction of red tape and bureaucracy.

### X. Concluding remarks and policy implications
- PICs face special geographic disadvantages constraining growth, but scope exists to raise growth by:
  - Increasing investment.
  - Promoting trade openness.
  - Ensuring political stability.
- Lower investment and exports in the Pacific are not entirely the result of geography, implying policy can make a difference.
- Large cross-country variation within the Pacific suggests policies matter; country-specific priorities and detailed case studies are desirable.
- Evidence suggests recent decades may have seen fundamental changes (policy settings or external environment) making size and location more determining factors for growth.
- A competitive exchange rate can be key to inducing higher investment and greater trade openness, but managing macroeconomic policies is necessary given size- and geography-induced cost disadvantages and large aid inflows.
- Maintaining competitiveness requires structural reforms to improve efficiency in production, transport, and communication; past reforms have produced payoffs (e.g., declines in telecommunication costs after deregulation).
- The current unfavorable global environment should strengthen resolve to accelerate high-payoff reforms.

### XI. Appendices — selected dataset indicators and country coverage
- Country list (Table A1) — Pacific (9): Fiji; Kiribati; Micronesia, Fed. Sts.; Palau; PNG; Samoa; Solomon Islands; Tonga; Vanuatu.
- Table A2: Average values over 1992-2008 (selected group comparisons; values reproduced exactly):
  - Convergence effect (lnY0): Small Economies: 7.98; Non-Pacific Small Economies: 8.15; Pacific Island Countries: 7.40.
  - Investment (ln(investment/GDP)): Small Economies: 3.26; Non-Pacific Small Economies: 3.33; Pacific Island Countries: 3.04.
  - Population growth (ln(n+g+d)): Small Economies: 1.88; Non-Pacific Small Economies: 1.88; Pacific Island Countries: 1.87.
  - Aid (aid/GNI, %): Small Economies: 7.96; Non-Pacific Small Economies: 5.69; Pacific Island Countries: 20.01.
  - Political stability: Small Economies: 0.49; Non-Pacific Small Economies: 0.53; Pacific Island Countries: 0.57.
  - Exports (in percent of GDP): Small Economies: 52.11; Non-Pacific Small Economies: 56.97; Pacific Island Countries: 35.08.
  - Volatility (growth rate standard deviation, %): Small Economies: 4.61; Non-Pacific Small Economies: 4.36; Pacific Island Countries: 5.48.
  - Per capita GDP growth (%): Small Economies: 2.38; Non-Pacific Small Economies: 2.87; Pacific Island Countries: 0.64.
  - GDP weighted distance to major overseas markets (ln): Small Economies: 9.01; Non-Pacific Small Economies: 8.93; Pacific Island Countries: 9.25.

*Italic: Source: _wp13104 - References (content excerpt provided).*

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

### _wp13104 - References

### I. INTRODUCTION
- Economic recovery in most Pacific island countries (PICs) following the global financial crisis has been weak.
- Two resource-rich countries, Papua New Guinea (PNG) and Solomon Islands, have rebounded strongly on the back of high commodity prices.
- Most other countries: annual GDP growth averaged just 0.3 percent during the period 2008-2010.
- At least six PICs out of the eleven IMF members have yet to reach their pre-crisis per capita GDP levels.
- Paper’s core questions:
  - To what extent has economic growth in PICs been slower than other small states and other similar country groups?
  - Have natural conditions (small size, geographical location) and external shocks (fuel prices, changing trade preferences) played a role in PICs’ long-term growth?
  - Why has growth in PICs slowed over the past decade?
  - How might economic and other policies have affected PICs’ growth performance?
- Scope: unique dataset for small states; long-run perspective; case studies; comparison with other small states and country groups.

### II. GROWTH PERFORMANCE—STYLIZED FACTS
- Historical growth comparisons:
  - 1970s: PIC economies grew at just below 4 percent per year.
  - 1990s: PIC growth rebounded and nearly caught up with LICs.
  - 2000s: sharp divergence; PICs had the weakest growth in four decades, managing to grow at only one-third the rate of LICs.
- Variation across PICs:
  - PNG and Solomon Islands: dramatic improvement in the last decade due to commodity booms.
  - Fiji: after outperforming neighbors historically, registered one of the slowest growth rates in the past decade largely due to domestic developments.
- Volatility:
  - Growth in PICs has been volatile but on average exhibited similar levels of volatility as other developing-country groups.
  - 2000s: PICs had lower growth volatility (along with the slowest growth rate) compared with earlier decades.

### III. GROWTH CONSTRAINTS IN SMALL STATES
- Theoretical disadvantages linked to small country size and geographical location:
  - Diseconomies of scale, difficulty exploiting economies of scale and industrial clustering.
  - Remoteness and insularity raise transportation and communication costs.
  - High unit costs of providing public goods because of indivisibility.
- Empirical literature:
  - Mixed evidence on smallness and remoteness effects: some studies find no systematic growth penalty for small states.
  - Greater trade openness often offsets disadvantages: small states’ trade-to-GDP ratio can be much higher (Easterly and Kraay (2000) sample: small states have trade-to-GDP 54 percentage points higher than average).
  - Regional/neighborhood effects: Gibson and Nero (2006) find that each percentage point increase in neighborhood average growth raises small-state GDP growth by 0.54 percentage point.
  - Political-metropolitan patron effects: Bertram (2003) estimates a $1 increase in patron per capita GDP raises island states’ per capita GDP by $0.30–$0.56.

### IV. MODEL, DATA AND METHODOLOGIES
- Model: growth-initial level (beta-convergence) model; core equation expressed in equations (1) and (2).
- Calibration: set g + δ = 5%.
- Data:
  - Cross-section dataset covering period 1992-2008 for 45 small states (some regressions cover 40 or 39 due to data limitations).
  - All money values in constant 2005 prices in US dollars.
- Variables used (dependent variable gy = average real growth of per capita GDP over 1992-2008):
  - ln(y0): GDP per capita in 1992 (captures convergence effect).
  - ln(investment rate): investment-to-GDP (%) averaged over 1992-2008.
  - ln(n + g + d): population growth plus g + δ (%) averaged over 1992-2008.
  - Aid/GNI: foreign aid-to-GNI ratio (%) — most recent data (2007) used.
  - Political stability: World Bank Worldwide Governance Indicators (averaged over available years 1996-2008).
  - Exports/GDP: exports-to-GDP ratio (%) averaged over 1992-2008.
  - Growth volatility: standard deviation of GDP growth rates over the period.
  - ln(distance): GDP-weighted distance to capital cities of major overseas markets.
  - Regional and special dummies (including Equatorial Guinea) to control heterogeneity.
- Insignificant variables tested (not robust): population size, education proxies, inflation, government expenditure, imports, agriculture/manufacturing shares, urban population ratio, arable land ratio, terms of trade, disaster indicators, trade-weighted external demand, external demand measured by top three export destinations’ GDP growth.
- Endogeneity and estimation:
  - Investment and aid tested for endogeneity; tests show no evidence of endogeneity for investment and aid in this sample (Sargan, Hausman tests).
  - Growth volatility found endogenous; political stability and exports-to-GDP used as instruments; two-stage least squares (2SLS) used where volatility is included.
  - Simple-to-general estimation strategy: start with neoclassical variables ln(y0), ln(s), ln(n + g + d), then add controls.

### V. DETERMINANTS OF GROWTH IN SMALL STATES — KEY FINDINGS
- Geography:
  - Regression results indicate geography has a large influence: PICs suffer a distance-related disadvantage in per capita GDP growth of about 1½ percentage points compared with an average non-Pacific small state.
  - PICs’ annual average per capita GDP growth over 1992-2008 was a little over 0.7 percent; without geographical disadvantage, PICs could have grown more than three times as fast during the period.
- Multivariate regression highlights (based on regressions 2 and 3 in Table 1):
  - Coefficients and t-stats reproduced as in source (selected values shown exactly as in Table 1):
    - Pacific constant: -3.13  [-3.33]
    - ln(y0): 7.00  [1.56]; in regression 2: -1.90  [-2.11]; in regression 3: -1.36  [2.90]
    - ln(invest. rate): -0.13  [-0.58]; regression 2: 15.68  [3.47]; regression 3: 1.08  [-4.77]
    - ln(n + g + d): 1.23  [2.11]; regression 2: -1.30  [-3.80]; regression 3: -3.94  [2.20]
    - Aid/GNI: -3.21  [-2.04]; regression 2: 1.05  [1.78]; regression 3: -0.06  [-3.32]
    - Political stability: -1.90  [-2.11]; regression 2: -0.05  [-2.61]; regression 3: -0.07  [-2.44]
    - Exports/GDP: 15.68  [3.47]; regression 2: 1.25  [2.50]; regression 3: 1.41  [3.19]
    - Growth volatility: -1.30  [-3.80]; regression 2: 0.03  [2.07]; regression 3: 0.04  [2.79]
    - ln(distance): 1.05  [2.11]; regression 2: 35.09  [2.90]; regression 3: -2.31  [-1.91]
  - Sample sizes and model fit:
    - Sample size: 45 (regression 1), 40 (regression 2), 39 (regression 3), 40 (2SLS regression 4).
    - Adjusted/centered R2: 0.8161 (regression 1), 0.8786 (regression 2), 0.9068 (regression 3), 0.8818 (regression 4).
    - Mean Variance Inflation Factor (VIF): 2.33 (regressions 1 and 2), 1.99 (regression 3).
    - Breusch-Pagan p-values: 0.3634 (regression 1), 0.4534 (regression 2), 0.7104 (regression 3).
    - Sargan statistic (regression 4): 0.7387.
- Interpretation:
  - Distance (remoteness) is a principal source of PICs’ growth disadvantage among small states.
  - Geography alone does not fully explain slower PIC growth; other significant determinants include exports openness, political stability, volatility, investment, and aid effects (as shown in regressions with varying signs and significance across specifications).
  - The impact and statistical significance of variables change across model specifications, indicating sensitivity to model choice and controls.

*Italic: Source: _wp13104 - References (content excerpt provided).*

### 1. y

### 1. y

### Methodology and regression diagnostics
- g + d = 0.05 is adopted in this study.
- Regressions 1~3 employ the ordinary least squares estimator while Regression 4 employs the two-stage least squares estimator.
- Growth volatility is found to be endogenous; political stability and exports/GDP are used as instrument variables for growth volatility.
- The validity of the instruments is confirmed by the Sargan statistic.
- Variance inflation factor (mean VIF) is used to detect collinearity; a mean VIF of less than 10 is taken as no evidence of collinearity problem.
- Breusch-Pagan test is employed to test for heteroskedasticity; a p-value greater than a preferred significance level is taken as no evidence of heteroskedasticity.
- All regressors are significant at least at the 10% level, with most of them significant at either 1% or 5% level.
- Dummy variables represent Equatorial Guinea and continents such as Africa, America, Europe and Pacific to help control for heteroskedasticity.

### Main determinants of PIC per capita GDP growth (relative to other small states)
- Initial per capita income in 1992 (convergence effect)
  - PICs were on average poorer than other small states in the early 1990s; this lower initial income allowed PICs to grow about ½ percentage point faster than an average non-Pacific small state.
- Investment (gross fixed capital formation out of GDP)
  - Over 1992-2008, investment in PICs averaged 22½ percent of GDP, about 6 percentage points lower than the average of all small states.
  - Had PICs achieved the average investment rate of non-Pacific small states, their real per capita GDP growth would have been about ¼ percentage point higher.
  - Large variations across PICs: Fiji’s investment rate is 14 percent of GDP while that of Vanuatu is 20 percent; this translates to a difference in the growth rate of ¼ percentage point.
- Exports-to-GDP (export openness)
  - On average each 10 percentage point increase in the exports-to-GDP ratio raises per capita GDP growth by about 0.3 percentage point.
  - The exports-to-GDP ratio in PICs is 24 percentage points lower than the average of non-Pacific small states, implying PICs could have grown by 0.6 percentage point faster had they exported as much as other small states (in percent of GDP).
  - Greater export openness increases output volatility: each 10 percentage point change in export openness is associated with 0.2 percentage point change in the standard deviation of GDP growth.
  - Less than a quarter of the growth benefits from greater export openness is nullified by the associated increase in output volatility, leaving PICs worse off by about 0.5 percentage point from their lower openness compared with other small states.
- Imports and trade turnover
  - Greater imports as percent of GDP have no statistically significant impact on growth.
  - Openness measured by the ratio of trade turnover (exports plus imports) to GDP shows no positive impact on growth in the estimations reported.
- Aid
  - Aid is found to be associated with slower GDP growth: for each 10 percentage point increase in aid as percent of GDP, growth is lower by 0.6 percentage point.
  - Preliminary tests indicate no reverse causality (slower growth does not lead to more aid); the estimated relationship applies to total aid (no disaggregation by type available).
  - Negative relationship should not be interpreted as aid lowering welfare; much aid is aimed at reducing poverty rather than increasing economic growth.
- Political stability
  - Measured by the World Bank Political Stability index, PICs score more favorably than other small states.
  - This gives PICs, on average, an advantage of 0.3 percentage point in growth over other small states.
  - Political stability varies substantially among PICs; other things being equal, the highest-scoring country in the region has a growth advantage of more than one percentage point over the lowest-scoring country.
- Growth volatility
  - Growth volatility is measured by standard deviation of per capita GDP growth rate.
  - Impacts of political stability and exports-to-GDP operate both directly on growth and indirectly via affecting growth volatility.

### Decomposition and quantitative magnitudes (summary of relative impacts)
- Combined policy- and geography-related factors, together with other determinants, collectively lower the per capita GDP growth in PICs by about another percentage point compared with an average small state (as indicated in the text discussion and Figure 5).
- Specific magnitude highlights:
  - Initial income: +½ percentage point advantage for PICs (convergence effect).
  - Investment shortfall: −¼ percentage point from lower investment (if raised to non-Pacific small states’ average).
  - Exports shortfall: −0.6 percentage point from being 24 percentage points lower in exports-to-GDP.
  - Export-induced volatility: each 10 percentage point increase in export openness raises standard deviation by 0.2 percentage point (less than a quarter of export benefits offset).
  - Aid effect: −0.6 percentage point for each 10 percentage point increase in aid/GDP.
  - Political stability advantage: +0.3 percentage point on average across PICs.

### Explaining the growth slowdown in the 2000s (2001-08)
- Period split and findings
  - The sample period 1992-2008 is divided into 1992-2000 and 2001-08 to assess the contribution of determinants to the growth slowdown.
  - Statistical tests suggest no significant structural breaks in growth for the entire sample of small states; the same regression coefficients are used to predict growth in sub-periods.
- Key contributor to slowdown
  - A decline in the exports-to-GDP ratio is highlighted as a major contributor to the growth slowdown in the 2000s.
  - The average export openness ratio fell by as much as 4½ percentage points in the 2000s from its 1990s level.
  - Investment increased slightly and output volatility declined, which together should have added over half percentage point to growth, but the decline in export openness offset part of these gains.
- Competitiveness and REER
  - Real exchange rate appreciation may have weakened competitiveness in several countries.
  - In the 2000s the REER appreciated for all six countries with their own currencies except Fiji (figure references in the source).

### Caveats, data limitations, and interpretation
- Results are generally consistent across model specifications but subject to data limitations, especially for small states.
- Sample size considerations and statistical significance caveats:
  - The example linking export openness to volatility was noted as illustrative only: estimates significant only at 83 percent confidence level and sample size small (28 observations).
- Policy implications must be guided by country-specific economic conditions:
  - Greater export openness is beneficial for growth, but may not be equally feasible for micro-states with limited export capacity.
  - Scaling up investment requires identifying impediments to investors and projects with high social returns; debt sustainability and capacity constraints must be considered.
- Human capital
  - Secondary education was included in regressions but found statistically insignificant in this analysis.

*Source: _wp13104 - 1. y.*

### Box 1. Samoa: Managing Shocks and Regaining Growth Momentum

### Box 1. Samoa: Managing Shocks and Regaining Growth Momentum

### Background and shocks
- Samoa experienced strong growth until the mid 2000s, supported by well coordinated reforms in the mid-1990s and sound macroeconomic policies that delivered an annual average GDP growth rate of 5 percent.
- Investment related to the 2007 Pacific Games provided an additional boost.
- A sequence of shocks then hit the economy: the food and fuel price hikes, a devastating tsunami (2009), and the global financial crisis (2008).

### Growth and fiscal outcomes
- Despite a large boost to government expenditure in FY2009/10 - 2011/2012, economic growth slowed significantly.
- The average overall fiscal deficit during this period was 7.3 percent of GDP.
- Real GDP growth between 2006/07 to 2010/11 averaged only 0.3 percent, the slowest since the mid 1960s.

### Public debt and external sectors
- The post-2009 tsunami infrastructure rebuilding and reconstruction was well supported by development partners through grants and concessional loans.
- Samoa’s public debt continued to rise, reaching well over 50 percent of GDP at the end of FY 2010/11.
- Key domestic sectors have shown weak performance:
  - Tourism and remittances have recovered only slowly.
  - The Yazaki automotive component plant has been declining since the global financial crisis erupted in 2008.
  - Agriculture has been stagnating in recent years.
- The increase in the trade deficit was largely due to increases in imports as a result of post-tsunami reconstruction; the strong Tala may also have helped boost imports in general.

### Competitiveness, policy challenges, and recommendations
- Samoa faces a difficult path to wind down fiscal deficits while the private sector has not yet taken up the slack.
- The government is committed to bringing down public debt to a more sustainable level over the medium term to maintain macroeconomic stability.
- While fiscal consolidation is essential for long-term sustainability, it may add further headwinds to an already weak recovery.
- Given considerable real exchange rate appreciation over recent years and weak global demand, substantial efforts are needed to:
  - Strengthen competitiveness.
  - Revitalize private sector-led growth.
- A re-assessment of the economy’s underlying strength and growth potential may be needed to establish a macroeconomic framework to regain competitiveness and maintain macroeconomic stability over the medium term.
- Structural reform needs to progress further to provide a more favorable environment for the private sector to take a lead role in economic development.
- Samoa’s strong track record of reform and the government’s continued commitment to keeping this record bodes well for future progress.

### Key statistics and facts (verbatim from source)
- Annual average GDP growth rate in mid-1990s period: 5 percent.
- Government expenditure boost period: FY2009/10 - 2011/2012.
- Average overall fiscal deficit: 7.3 percent of GDP.
- Real GDP growth, 2006/07 to 2010/11: averaged 0.3 percent.
- Slowest growth since: the mid 1960s.
- Public debt at end of FY 2010/11: well over 50 percent of GDP.
- Notable shock events: food and fuel price hikes; a devastating tsunami (2009); the global financial crisis (2008).
- Declining firm: Yazaki automotive component plant (declining since 2008).

*Source: IMF staff box “Samoa: Managing Shocks and Regaining Growth Momentum”*

### introduction of jet engines in the late 1950s and containerized shipping in the 1960s have

### _wp13104 - introduction of jet engines in the late 1950s and containerized shipping in the 1960s have

### Transportation developments and impacts on PICs
- Introduction of jet engines in the late 1950s and containerized shipping in the 1960s substantially increased the efficiency of air and ocean transportation over past decades.
- Despite technological advances, ocean transport costs (as percent of the values of shipped goods) in the early 2000s were reported to be as they were in the 1950s (Hummels, 2007).
- Rapid increases in transportation cost over the past decade, especially for air transportation, led to:
  - A sharp decline in the proportion of airlifted goods in trade after a long period of steady increase (Figures 14 and 15).
  - For PICs, higher air transport costs have added substantially to the transport cost of perishable products (fish, vegetables and fruits).
- Containerized shipping was introduced in PICs mostly in the late 1970s and 1980s, later than elsewhere, due to substantial investment needs (including port facilities).
- Competition in PICs’ air and ocean transport industries is limited because of small market size and state monopoly; benefits of technological advances may not have passed onto consumers as much as in other parts of the world.
- Technological advances produced faster speed and greater reliability in shipping, which in principle could lower transport cost relative to competitors; however:
  - Remote, small states such as PICs cannot take much advantage of large, more efficient vessels because of small trade volumes and the need to stop more frequently to serve small and dispersed destinations.
  - Advances have tended to favor exporters of manufactured goods because large vessels and greater speed/reliability allow finer segmentation of value chains.
  - As the spread between high-priced and low-priced goods in each product category widens, the cost advantage enjoyed by high-end goods is growing over time (Hummels, 2007; 2009).
  - Without high value-added exports, PICs could be increasingly disadvantaged by transport cost.

### Key empirical findings relevant to growth and aid
- Panel regression result cited: A 10 percent increase in tourist arrivals per capita raises economic growth by about 0.2 percent (Thacker, 2011).
- Weight-based measures tend to overstate the decline in transport cost because composition of shipped goods shifts toward lighter, higher-valued goods over time.
- Trade partners with 10 percent more exports enjoy 0.8 percent lower transport prices (Kleinert and Spies, 2011).
- Lack of competition in telecommunications raised costs to consumers in Samoa (Favaro et al., 2008).

### Policy implications: macroeconomic management
- Remoteness imposes a lower speed limit on PICs' economic growth and contributes to higher growth volatility; policymakers should set realistic growth targets and maintain larger policy buffers in good times.
- Small, relatively poor countries often receive more aid (relative to GDP) than large countries, reducing the need for a depreciated real exchange rate to balance trade or current account; this can raise living standards via better public services and cheaper imports but places pressure on export competitiveness.
- Recommendations to alleviate adverse effects of aid-supported strong exchange rates:
  - Direct more aid to productivity improvements.
  - Accelerate structural reforms.
- Exchange rate regimes:
  - Both fixed and flexible regimes have trade-offs in PICs.
  - Fixed pegs provide a useful nominal anchor but can become unsustainable if inflation runs persistently higher than trading partners, necessitating large devaluations.
  - Flexible regimes can lead to rapid appreciation when there are large foreign exchange inflows (aid, resource rents), which can be sterilized through reserve accumulation but sterilization is often incomplete due to costs and higher domestic interest rates.
  - Foreign exchange markets are shallow; a floating regime may cause large exchange rate volatility harmful for trade.
- Macroeconomic coordination recommendations:
  - Under a fixed exchange rate, monetary policy must be subordinate to exchange rate policy and ensure inflation is not consistently higher than partner countries.
  - Fiscal policy must be prudent to avoid pressure on domestic price levels and debt sustainability.
  - Large windfall inflows could be saved (e.g., through a sovereign wealth fund) for inter-generational distribution and consumption smoothing.
- Since the global financial crisis, PIC macroeconomic policies have been accommodative (both monetary and fiscal) to support growth; however, higher inflation from rising world food and fuel prices has led to significant real exchange rate appreciation in a number of countries.
- As economic recovery strengthens, import demand increases, and aid inflows decline, PICs may face challenges reverting real exchange rates to more sustainable and competitive levels.

### Structural reforms, competitiveness, and integration
- To mitigate the effects of strong exchange rates, raising productivity faster than trading partners is essential when facing persistently higher inflation under a fixed exchange rate.
- Empirical constraint: Productivity growth in PICs has been generally slow; low investment is a key constraint on higher growth.
  - Example: Total factor productivity growth in Fiji over 1983–2007 was estimated at only 0.5 percent per year, less than half the rates in Asian and Pacific countries (Chen and Singh, 2011).
- Political stability is important for growth in PICs and should reinforce reform efforts toward better governance.
- Reducing the cost of distance and insularity is critical for competitiveness:
  - Transportation and communication industries exhibit economies of scale, disadvantaging small and remote countries.
  - Challenges include scaling up investment in transport and communication infrastructure and implementing reforms and regulation to ensure rigorous competition.
  - Greater regional cooperation (e.g., Pacific Forum Line; Pacific Islands Telecommunications Association) can help mitigate diseconomies of scale.
- Trade policy and global integration:
  - Further integration with the global economy is the ultimate way to overcome smallness and distance.
  - PICs should seek deeper integration with metropolitan countries, particularly in trade in labor services.
  - Temporary migrant worker programs with New Zealand and Australia have progressed, but potential remains large relative to PIC labor forces.
  - PACER-Plus provides a framework for further integration with Australia and New Zealand; accelerating negotiations could be beneficial.
  - Strengthening trade and financial ties with Asia is important:
    - PICs without mineral resources have not fully taken advantage of increasingly affluent Asian markets due to supply constraints and need for reliable supply and marketing.
    - Some emerging Asian economies offer duty and quota-free entry of goods from UN-defined LDCs among PICs.
    - Secure market access in Asia could encourage foreign investment and increase awareness of Pacific goods; Asian entrepreneurs could bring capital, skills, and marketing.

### Box 3 — Fiji: lessons and hopes (key statistics and findings)
- Fiji averaged real GDP growth of 2.1 percent over the past two decades.
- Since 1987, Fiji experienced four coups; research shows coups had an immediate negative effect on the economy.
- Private investment has been declining and volatile since 1988.
- Public debt reached 51.5 percent of GDP.
- A 2009 devaluation boosted the export sector, especially tourism, and helped restore macroeconomic stability.
- The economy grew 2 percent in 2011.
- Total factor productivity growth in Fiji over 1983–2007 estimated at 0.5 percent per year (Chen and Singh, 2011).
- Structural issues highlighted:
  - Declining public investment and maintenance of infrastructure until recently.
  - High and persistent inflation relative to trading partners under a pegged exchange rate eroded competitiveness.
  - Uncertainty over renewal of land leases reduced investment and productivity in land-based industries.
  - Mismanagement of the Fiji Sugar Corporation accelerated decline of the sugar industry and strained public finances.
  - Reforms undertaken include creation of a land bank, relaxation of foreign exchange controls, and reduction of red tape and bureaucracy.

### Concluding remarks
- PICs face special geographic disadvantages constraining growth, but there is scope to raise growth by increasing investment, promoting trade openness, and ensuring political stability.
- Lower investment and exports in the Pacific are not entirely the result of geography, implying policy can make a difference.
- Large cross-country variation within the Pacific suggests policies matter; country-specific priorities and detailed case studies are desirable.
- Evidence suggests recent decades may have seen fundamental changes (policy settings or external environment) making size and location more determining factors for growth.
- A competitive exchange rate can be key to inducing higher investment and greater trade openness, but managing macroeconomic policies is necessary given size- and geography-induced cost disadvantages and large aid inflows.
- Maintaining competitiveness requires structural reforms to improve efficiency in production, transport, and communication; notable past payoffs include marked declines in telecommunication costs following deregulation.
- The current unfavorable global environment should strengthen resolve to accelerate high-payoff reforms.

*Italic source attribution: IMF Working Paper content unit _wp13104*

### References

### _wp13104 - References

### Bibliographic coverage
- Extensive citations covering empirical growth, volatility, trade costs, small states, aid, and geography.
- Key recurring topics in the references:
  - Volatility and investment in developing countries (Aizenman and Marion, 1999; Ramey and Ramey, 1995; Malik and Temple, 2009).
  - Aid and growth in Pacific/island contexts (Ali and Isse, 2005; Pavlov and Sugden, 2006; Rao et al., 2008; Feeny and McGillivray).
  - Trade costs, transportation and geography (Anderson and van Wincoop, 2004; Hummels, 2007; Hummels, 2009; Gallup, Sachs, and Mellinger, 1998; Kleinert and Spies, 2011; UNCTAD, 2010).
  - Small states and micro-state economic performance, vulnerabilities, and policy challenges (Armstrong and Read, multiple works; Briguiglio, 1995; Commonwealth Secretariat and World Bank, 2000; Streeten, 1993; Srinivasan, 1986).
  - Growth diagnostics and productivity (Hausmann, Rodrik, and Velasco, 2005; Mankiw, Romer, and Weil, 1992; Islam, 2003a/b; Narayan et al., 2006).
  - Region- and country-specific studies for Pacific Island Countries (Chen and Singh, 2011; Gibson and Nero, 2007; Singh, 2005; Yang et al., 2011).
- Methodological and theoretical foundations cited include cross-country growth regressions, Hodrick-Prescott filtering, transport cost modeling, and macroeconometric modeling.

### Appendix I — Growth Trends in PICs (Hodrick-Prescott Filter lambda=100)
- Series of country-level growth trend and cycle illustrations using the Hodrick-Prescott Filter (lambda=100) for:
  - PNG (labels: TrendCycle, PNG)
  - Solomon Islands (labels: TrendCycle, Solomon Islands)
  - Vanuatu (labels: TrendCycle, Vanuatu)
  - Fiji (labels: TrendCycle, Fiji)
  - Samoa (labels: TrendCycle, Samoa)
  - Tonga (labels: TrendCycle, Tonga) — appears twice in the listing
  - Marshall Islands (labels: TrendCycle, MARSHALL ISLANDS)
  - Palau (labels: TrendCycle, PALAU)
  - Kiribati (labels: TrendCycle, Kiribati)
  - Micronesia, Fed. States (labels: TrendCycle, MICRONESIA, FED. STATES)
- Charts include axis ranges shown in the source (examples):
  - Value ranges illustrated: -20, -15, -12, -10, -8, -6, -5, -4, -2, 0, 2, 4, 5, 6, 8, 10, 12, 15, 20, 30, 40, 50 (as they appear across different country panels).
- The Appendix I pages contain visual time series for 1965–? (x-axis tick labels shown as 65 70 75 80 85 90 95 00 05 10 15 20 in compact form), indicating long-run trend and cyclical components by country.

### Appendix II — Country List and Data Description
- Table A1: Country List categorized by region
  - Africa (14): Botswana; Cape Verde; Comoros; Djibouti; Equatorial Guinea; Gabon; Gambia, The; Guinea-Bissau; Lesotho; Mauritius; Namibia; Sao Tome and Principe; Seychelles; Swaziland.
  - America (12): Antigua and Barbuda; Bahamas, The; Barbados; Bermuda; Belize; Bermuda (duplicated in formatting of source); Dominica; Grenada; Guyana; St. Kitts and Nevis; St. Lucia; St. Vincent and the Grenadines; Suriname. (Note: source layout groups countries in multiple columns; entries reproduced as they appear.)
  - Europe (4): Estonia; Malta; Montenegro; San Marino.
  - Pacific (9): Fiji; Kiribati; Micronesia, Fed. Sts.; Palau; PNG; Samoa; Solomon Islands; Tonga; Vanuatu.
  - Asia (6): Bahrain; Bhutan; Brunei Darussalam; Maldives; Qatar; Timor-Leste.
- Table A2: Average Values of Relevant Series over 1992-2008 (presented for four groups: Small Economies; Non-Pacific Small Economies; Pacific Island Countries)
  - Convergence effect (lnY0)
    - Small Economies: 7.98
    - Non-Pacific Small Economies: 8.15
    - Pacific Island Countries: 7.40
  - Investment (ln(investment/GDP))
    - Small Economies: 3.26
    - Non-Pacific Small Economies: 3.33
    - Pacific Island Countries: 3.04
  - Population growth (ln(n+g+d))
    - Small Economies: 1.88
    - Non-Pacific Small Economies: 1.88
    - Pacific Island Countries: 1.87
  - Aid (aid/GNI, %)
    - Small Economies: 7.96
    - Non-Pacific Small Economies: 5.69
    - Pacific Island Countries: 20.01
  - Political stability
    - Small Economies: 0.49
    - Non-Pacific Small Economies: 0.53
    - Pacific Island Countries: 0.57
  - Exports (in percent of GDP)
    - Small Economies: 52.11
    - Non-Pacific Small Economies: 56.97
    - Pacific Island Countries: 35.08
  - Volatility (growth rate standard deviation, %)
    - Small Economies: 4.61
    - Non-Pacific Small Economies: 4.36
    - Pacific Island Countries: 5.48
  - Per capita GDP growth (%)
    - Small Economies: 2.38
    - Non-Pacific Small Economies: 2.87
    - Pacific Island Countries: 0.64
  - GDP weighted distance to major overseas markets (ln)
    - Small Economies: 9.01
    - Non-Pacific Small Economies: 8.93
    - Pacific Island Countries: 9.25

*Source: _wp13104 - References (source PDF content).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13104.pdf_