## Annex V. Vertical Policy Tools, Scope and Limitations for PICs

## Source details

**Canonical URL:** [Annex V. Vertical Policy Tools, Scope and Limitations for PICs](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025111-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025111-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025111-print-pdf.pdf.json)

---

### I. Introduction and context
- Time horizon for empirical analysis: 1990 until 2019, using non-overlapping three-year periods.
- Sample: about 180 countries, including 11 of the 12 PICs.
- Comparator benchmark used in counterfactuals: average annualized 3-year growth rate of other Asian LICs between 1999 and 2019, equal to 4.3 percent.
- Key structural constraints identified for PICs:
  - Small and isolated domestic markets (small populations, extreme geographic dispersion).
  - Higher transportation and raw material costs due to remoteness.
  - High vulnerability to climate change and frequent, rising-severity natural disasters.
  - Dependence on tourism and natural resources; reliance on food and capital imports and official aid.
  - Some PICs face weaknesses in macroeconomic policy frameworks and tight capacity constraints leading to weak policy implementation and/or unsustainable exploitation of natural resources.

### II. Empirical approach to tourism and growth
- Dependent variable: annualized growth rate of per capita GDP in constant 2011 PPP dollars for country i at period t.
- Core regression controls: tourism specialization, initial per capita GDP, trade openness, net FDI flows, number of natural disasters, country fixed effects (γi) and period fixed effects (δt).
- Instruments used to address endogeneity of tourism:
  - Lagged value of the tourism average variable.
  - Period-start value of the tourism variable.
  - Average tourism variable of each country’s neighbors.
- Three measures of tourism specialization employed:
  - International tourism receipts as percentage of total exports.
  - Logarithm of level of international tourism receipts.
  - Logarithm of number of international tourist arrivals.
- Estimation methods: OLS for baseline and interaction specifications; IV estimation for baseline with instruments above.

### III. Findings on tourism-led strategies
- Regression findings:
  - Tourism variables are positively correlated with growth in general; interaction term for PICs is insignificant.
  - On margins: a 1 percent increase in tourist arrivals (extensive margin) has a larger impact on growth than a 1 percent increase in tourism receipts (intensive margin).
  - Some IV regressions for specific tourism variables yield coefficients not statistically different from zero.
- Scaling-up required to match comparator growth:
  - To reach the 4.3 percent benchmark, PICs would need on average to more than double their tourism exports share.
  - For tourism-specialized PICs (Fiji, Palau, Tonga, Vanuatu, and Samoa), even a tourism export share of 100 percent would not suffice to reach the 4.3 percent average.
- Policy-relevant implications:
  - Focus on increasing the exports share of tourism by increasing international tourism relative to domestic tourism and by increasing domestic value added in the tourism sector.
  - Recognize limits to tourism specialization; sizable investments in infrastructure, connectivity, and marketing are needed where scaling-up is required.

### IV. Measuring diversification and identifying episodes
- Diversification dimensions:
  - Export goods and services diversification.
  - Trading partners diversification.
  - Domestic production diversification.
- Diversification metric: Theil index defined as Theil = (1/N) * Σ (x_s / μ) * ln(x_s / μ); a decrease indicates increased diversification.
- Data sources:
  - Exports diversification index: COMTRADE goods trade data.
  - Domestic diversification: UN value added data (10 sectors) and WB WDI 3-sector data.
- Episode detection algorithm:
  - Apply HP filter with smoothing parameter equal to 100 for annual data.
  - Estimate number of breaks using Bai and Perron (1998) with trimming set to 0.15.
  - A diversification episode starts with a downward break in the filtered trend; episode ends upon an upward break, a small downbreak (<5 percent), reaching a valley, or end of sample.

### V. Summary statistics on diversification episodes
- Aggregate counts and durations (sample periods 1962–2019, Table 4):
  - PICs: 26 total episodes; average duration per episode: 4.5 years; 18 exports episodes, 8 domestic episodes.
  - Caribbean Small States: 33 total episodes; average duration 5.5 years; 24 exports episodes, 9 domestic episodes.
  - Other SIDS: 20 total episodes; average duration 3.9 years; 11 exports episodes, 9 domestic episodes.
  - Asian LICs (excluding PICs): 20 total episodes; average duration 7.6 years; 10 exports episodes, 10 domestic episodes.
- Timing and persistence:
  - PICs started diversification efforts later than comparators.
  - PIC episodes are, on average, shorter than comparators and more frequently reversed.

### VI. Quantitative case studies (Synthetic Control Method) — selected empirical results
- SCM outcome variables: real GDP per capita, output volatility (rolling 3-year standard deviation), and imports share.
- SCM pre-treatment matching: average number of natural disasters and lags of the outcome variable; use 10 pre-treatment years where possible; 5 post-treatment years constant.
- Inference: permutation (placebo) method; fit measures include Pre-RMSPE, Fit index, Relative fit index.

Selected case-study quantitative findings preserved verbatim:
- Fiji, exports diversification episode, 2001–2007:
  - Exports diversification index improved by 9.5 percent (declined from 3.49 to 3.16).
  - Effect on GDP peaks 2 years after the event, at about 0.4 percent of GDP per one percent improvement in diversification.
  - In-space placebo dynamic p-values: for GDP, in the first two years after the start of the event, about 40 percent of placebos result in an effect at least as large; for volatility, at least 50 percent of placebos result in an effect at least as large; for imports, only in the first year 20 percent of placebos have an effect at least as large.
- Samoa, exports diversification episode, 2008–2014:
  - Exports diversification index improved by 14.8 percent (from 3.98 to 3.39).
  - Average GDP improvement: 0.2 percent of GDP per one percent improvement in the exports diversification index.
  - Dynamic p-values: lie between 60-80 percent, indicating the measured effect is likely to be random.
  - Imports/GDP: sustained reduction of about 5 percentage points, or on average about 1 percent per one percent improvement in diversification.
- Marshall Islands, domestic diversification episode, 2010–2013:
  - Domestic diversification improved by 29.5 percent using the 3-sector index and 8.9 percent using the value-added index.
  - After a small positive effect at the start year, actual Marshall Islands had lower GDP than the counterfactual; estimated effect likely insignificant with p-values in the 60-90 percent range.
- Samoa, domestic specialization episode, 2011–2019:
  - Domestic diversification index increased from 0.29 to 0.34, a worsening in diversification of 20.2 percent.
  - Effect bottoms out 3 years after start at about -0.6 percent per percentage worsening in diversification.
  - Volatility: significant reduction estimated during years 3-4 after the start.

- Annex III fit examples (selected Pre-RMSPE and fit indices):
  - Fiji, 2001 — GDP: Pre-RMSPE: 154.4978; Fit index: 0.0158; Relative fit: 0.6842. Model: Average natural disasters. 10 pre-years.
  - Samoa, 2008 — GDP: Pre-RMSPE: 84.6961; Fit index: 0.0177; Relative fit: 0.6875. Model: Average natural disasters. 10 pre-years.
  - Marshall Islands, 2010 — GDP: Pre-RMSPE: 88.1608; Fit index: 0.0276; Relative fit: 0.6471. Model: Average natural disasters. 10 pre-years.

### VII. Implications for strategy design and policy recommendations
- Core empirical conclusions:
  - Tourism-led growth can contribute positively to growth, with stronger effects from the extensive margin (tourist arrivals) than the intensive margin (receipts per tourist) on average.
  - For many PICs the required expansion of tourism exports to match comparator growth is large and may be infeasible; tourism-specialized PICs face limits to further growth from tourism alone.
  - Export diversification episodes can produce measurable gains in output and reductions in volatility; domestic diversification has shown weaker effects in the PIC case studies examined.
  - Building resilience to climate change is a pre-condition for successful diversification-based strategies because damage and reconstruction from natural disasters disrupt diversification efforts.
- Recommended prioritization:
  - Continue and prioritize horizontal policies (education, health, infrastructure) to build human and physical capital and resilience to natural disasters.
  - Targeted vertical policies that diversify exports are likely to deliver the most growth impact and should complement horizontal policies.
  - Set up coordinated action plans with clear measurement, evaluation, and exit strategies to avoid past pitfalls of diversification efforts.
  - Promote joint and regional action to overcome market failures linked to geography and small domestic markets.
- Roadmap principles for implementing vertical policies (adapted from Cherif and others (2022)):
  1. Identify the externality the vertical policy aims to address; target sectors with higher distortions where potential gains are larger; consider regional initiatives to expand market scale and share fiscal burden.
  2. Ensure interventions pass a cost-benefit analysis including social benefits, costs, spillovers, volatility reduction, and distributional implications.
  3. Account for government failures: anticipate rent-seeking, corruption, and mercantilist pressures; implement monitoring and strengthen institutional capacity.
  4. Balance addressing market failures while mitigating government failure risks: include performance targets and exit strategies, maintain competition, focus on exports potential, and strengthen governance.
- Practical design considerations for PICs:
  - Prioritize instruments with clear fixed costs, time horizons, and measurable outcomes given narrow tax bases.
  - Favor regional cooperation (for example, FFA, PNA) to share monitoring, knowledge, and market development responsibilities.
  - Use regulatory sandboxes and enabling regulatory frameworks to foster technology-enabled innovation (example: Solomon Islands regulatory sandbox for financial inclusion).
  - Consider innovative domestic financing approaches (for example, National Provident Funds acting as or funding venture capital firms) with careful risk limits.

### VIII. Illustrative application: Fisheries and the Blue Economy (policy sequencing)
- Step 1: Market failures and externalities to address include the tragedy of the commons, learning-by-doing externalities, technology gaps, and connectivity constraints.
- Step 2: Sector-specific policy failures to assess include trade policy barriers, limited climate-proofed infrastructure, regulation of fishing practices, scarce specialized skills, and monitoring/data gaps.
- Step 3: Coordinated action plan examples:
  - Promote climate-proofed infrastructure; instruments could include tax holidays or credit-guarantees against infrastructure updates.
  - Enhance role of regional bodies (FFA, PNA) in cooperation, knowledge sharing, vessel monitoring, and export market development.
  - Perform regular HIES to evaluate distributional effects and progress.
  - Create regional incubators and mobile training institutes to improve post-harvest methods and diversify coastal fisheries.
  - Improve transparency of fisheries-related subsidies and implement exit strategies.
  - Continue horizontal policies that improve infrastructure, access to electricity, governance, and transparency; streamline regulations.
- Step 4: Define evaluation methods before implementation and use objective market signals (production and exports growth) to adapt policy tools.

### IX. Suitability and limitations of vertical policy tools for PICs (selected tool highlights)
- Product Market — Trade Policy:
  - Can promote exports but may lead to net welfare losses, limit technology adoption if applied to capital/intermediate goods, and entail large fiscal cost in a limited tax base environment.
  - Regional agreements can virtually increase domestic market scale.
- Product Market — Tax incentives:
  - Can attract FDI but risk tax base erosion if not time-bound; effectiveness depends on complementary policies.
- Product Market — SOEs:
  - SOEs exist in key infrastructure sectors; improving transparency, governance, and independent mandates is recommended.
- Capital Market — Directed lending and credit guarantees:
  - Can boost production if targeted firms face financing constraints; present fiscal risks and contingent liabilities.
- Capital Market — Venture capital and incubators:
  - Potential role for National Provident Funds when portfolio diversification is needed.
- Labor Market — Skills development:
  - Require strengthening primary and secondary education to be effective.
- Land Market — Cheap land and SEZs:
  - Can attract investment but raise concerns about land rights and potential to reduce incentives for economy-wide reforms.
- Technology — R&D incentives:
  - Returns tend to be smaller without adequate human capital; technology-transfer instruments and regional training institutes may be more effective.

### X. Conclusion — substantive findings and policy implications
- No single growth strategy fits all PICs; country-specific factors and constraints matter.
- Empirical synthesis:
  - For some PICs, a tourism-led strategy could raise growth if centered on increasing value-added and the exports share of tourism.
  - For most PICs, tourism-led strategies have limitations due to large investment needs or prior tourism-specialization; diversification-based strategies may offer larger benefits.
  - Quantitative case studies indicate exports diversification strategies have the most impact on lifting growth and reducing volatility.
  - Little evidence that diversifying only domestic output yields favorable growth impacts, though it may reduce output volatility and improve outcomes not analyzed (for example, food security).
  - Building resilience to climate change is a pre-condition for successful diversification-based growth strategies.
- Policy recommendations:
  - Continue and prioritize horizontal policies (education, health, infrastructure) and resilience investments.
  - Pursue targeted, well-designed vertical policies focused on export diversification, accompanied by governance improvements, cost-benefit analysis, monitoring, evaluation, and exit strategies.
  - Utilize regional cooperation to mitigate scale and fiscal constraints.

*Italic: Source — Annex V. Vertical Policy Tools, Scope and Limitations for PICs, from Growth Strategies and Diversification in the Pacific Islands Countries (IMF Working Paper).*

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

### References ................................................................................................................................................... 29

### Annexes and Sections

- Annex I. Country samples. ........................................................................................................................... 31
- Annex II. Diversification Episodes. ............................................................................................................... 32
- Annex III. Synthetic Control Method. ............................................................................................................ 38
- Annex IV. Exports Partner Diversification Episodes ..................................................................................... 41

*Source: wpiea2025111-print-pdf - References .............................................................................................................*

### Annex V. Vertical Policy Tools, Scope and Limitations for PICs .........................................................

### Annex V. Vertical Policy Tools, Scope and Limitations for PICs

### I. Introduction and context
- The analysis focuses on growth strategies and diversification in the Pacific Islands Countries (PICs), emphasizing tourism-led strategies and diversification as alternative or complementary paths.
- Time horizon for empirical analysis: 1990 until 2019, using non-overlapping three-year periods.
- Sample: about 180 countries, including 11 of the 12 PICs.
- Comparator benchmark used in counterfactuals: average annualized 3-year growth rate of other Asian LICs between 1999 and 2019, equal to 4.3 percent.
- Key structural constraints identified for PICs:
  - Small and isolated domestic markets (small populations, extreme geographic dispersion).
  - Higher transportation and raw material costs due to remoteness.
  - High vulnerability to climate change and frequent, rising-severity natural disasters.
  - Dependence on tourism and natural resources; reliance on food and capital imports and official aid.
  - Some PICs face weaknesses in macroeconomic policy frameworks and tight capacity constraints leading to weak policy implementation and/or unsustainable exploitation of natural resources.

### II. Empirical approach to tourism and growth
- Growth regression framework:
  - Dependent variable: annualized growth rate of per capita GDP in constant 2011 PPP dollars for country i at period t.
  - Core specification includes tourism specialization, controls (initial per capita GDP, trade openness, net FDI flows, number of natural disasters), country fixed effects (γi) and period fixed effects (δt).
  - Equations estimated: baseline (equation (1)) and interaction with PIC dummy (equation (2)).
- Instruments used to address endogeneity of tourism:
  - Lagged value of the tourism average variable.
  - Period-start value of the tourism variable.
  - Average tourism variable of each country’s neighbors (the latter instrument generally cannot be rejected as weak).
- Three measures of tourism specialization employed:
  - International tourism receipts as percentage of total exports (Table 1).
  - Logarithm of level of international tourism receipts (Table 2).
  - Logarithm of number of international tourist arrivals (Table 3).
- Estimation methods reported: OLS for equations (1) and (2); IV estimation for equation (1) with the instruments above.

### III. Findings on tourism-led strategies
- Regression results summary:
  - Tourism variables are positively correlated with growth in general; interaction term for PICs is insignificant, indicating PICs are not significantly different from other countries in the tourism–growth relation.
  - On margins: on average, a 1 percent increase in tourist arrivals (extensive margin) has a larger impact on growth than a 1 percent increase in tourism receipts (intensive margin).
  - In some IV regressions for specific tourism variables, coefficients are not statistically different from zero.
- Policy-relevant implications from results:
  - A tourism-led strategy should focus on increasing the exports share of tourism via:
    - Increasing international tourism relative to domestic tourism.
    - Increasing domestic value added in the tourism sector to raise tourism export value and build more inclusive value chains.
- Scaling-up required to match comparator growth:
  - To reach the 4.3 percent benchmark, PICs would need on average to more than double their tourism exports share.
  - For tourism-specialized PICs (Fiji, Palau, Tonga, Vanuatu, and Samoa), even a tourism export share of 100 percent would not suffice to reach the 4.3 percent average, indicating limits to tourism specialization alone.
  - Conclusion: tourism specialization can be a good fit for some PICs, but for many the required scaling-up is large and may be infeasible without sizable investments in infrastructure, connectivity, and marketing; tourism-specialized PICs will need additional strategies beyond further tourism specialization.

### IV. Measuring diversification and identifying episodes
- Diversification dimensions considered:
  - Export goods and services diversification.
  - Trading partners diversification.
  - Domestic production diversification.
- Diversification metric: Theil index of diversification defined as
  - Theil = (1/N) * Σ (x_s / μ) * ln(x_s / μ), where t indexes product/services categories or productive sectors, x_s is value share, and μ is the mean of x_s; a decrease in the index indicates an increase in diversification.
- Data and indices:
  - Exports diversification index computed using COMTRADE goods trade data.
  - Domestic diversification measured using UN value added data (10 sectors) and WB WDI 3-sector data; results presented as Theil indices (3-year moving averages).
- Stylized facts:
  - PICs on average lag comparators (Asian LICs and Caribbean small states) in diversification across measures.
  - Within PICs there is considerable heterogeneity; some PICs show consistent specialization in the late 2000s.
  - Domestic diversification for PICs tends to be stagnant or more specialized compared with comparators.
- Episode detection algorithm:
  - Apply Hodrick-Prescott (HP) filter to the diversification index to separate trend from cycle.
  - Estimate number of breaks using Bai and Perron (1998) methodology following Berg, Ostry and Zettelmeyer (2012).
  - A diversification episode starts with a downward break in the filtered trend; episode ends upon an upward break, a small downbreak (<5 percent), reaching a valley before trend reversal, or end of sample.
  - Annex II contains algorithm details and episode lists; Figures AII.1–AII.4 show episodes for PICs across dimensions.

### V. Summary statistics on diversification episodes
- Aggregate counts and durations (Table 4):
  - PICs: 26 total episodes across sample periods (1962–2019); average duration per episode: 4.5 years; 18 exports episodes, 8 domestic episodes.
  - Caribbean Small States: 33 total episodes; average duration 5.5 years; 24 exports episodes, 9 domestic episodes.
  - Other SIDS: 20 total episodes; average duration 3.9 years; 11 exports episodes, 9 domestic episodes.
  - Asian LICs (excluding PICs): 20 total episodes; average duration 7.6 years; 10 exports episodes, 10 domestic episodes.
- Timing:
  - PICs started diversification efforts later than comparators; Caribbean small states had more episodes starting in the 1990s; Asian LICs had many episodes in the 1970s.
  - PIC diversification episodes are, on average, shorter than comparators and more frequently reversed.

### VI. Quantitative case studies (synthetic control method) and main lessons
- Methodology:
  - For each non-reversed diversification episode, apply the synthetic control method (Abadie and Gardeazabal 2003; Abadie, Diamond and Hainmuller 2010) to build a synthetic counterfactual that replicates the pre-episode path.
  - Outcome variables analyzed: real GDP per capita, output volatility, and imports share.
  - Models fitted include average number of natural disasters as a covariate and lags of the outcome to preserve donor pool size (technical details in Annex III).
- Selected case studies (4 episodes with sufficient data and minimal overlap):
  - Exports diversification: Fiji (2001–07) and Samoa (2008–14).
  - Domestic diversification: Marshall Islands (2010–13) and Samoa (2011–10i).
- Main empirical lessons from SCM analysis:
  - Exports diversification has some potential to raise output and reduce output volatility.
  - Domestic diversification has had small to no impact on output and limited impact on volatility.
  - Domestic specialization episodes often follow natural disasters during reconstruction phases.

### VII. Implications for strategy design (summary points relevant to Annex focus)
- Tourism-led growth can contribute positively to growth, with stronger effects from the extensive margin (tourist arrivals) than the intensive margin (receipts per tourist) on average.
- For many PICs the required expansion of tourism exports to match comparator growth is large and may be infeasible; tourism-specialized PICs face clear limits to further growth from tourism alone.
- Export diversification episodes can produce measurable gains in output and reductions in volatility; domestic diversification has shown weaker effects in the PIC case studies examined.
- Timing and persistence matter: PICs’ later and shorter diversification episodes relative to comparators may have lasting adverse implications for growth performance.
- Designing growth strategies for PICs should account for:
  - Binding constraints from geography, market size, and disaster vulnerability.
  - The cost and feasibility of scaling up tourism (infrastructure, connectivity, advertising, domestic value added).
  - The potential returns to targeted export diversification and the shorter-lived nature of many PIC diversification episodes.

*Italic: Source — Annex V. Vertical Policy Tools, Scope and Limitations for PICs, from Growth Strategies and Diversification in the Pacific Islands Countries (IMF Working Paper). *

### Annex IV presents additional case studies considering exports partner diversification.

### Annex IV presents additional case studies considering exports partner diversification.

### Fiji, exports diversification episode, 2001-2007 — Only diversifying exports
- Diversification episodes identified for Fiji: 1969-1972, 1985-1991, and 2001-2007. Analysis focuses on 2001-2007.
- Exports diversification index: improved by 9.5 percent (the index declined from 3.49 to 3.16).
- GDP impact:
  - Actual Fiji has higher GDP compared with synthetic Fiji after the diversification episode.
  - Effect on GDP peaks 2 years after the event, at about 0.4 percent of GDP per one percent improvement in diversification.
- GDP volatility: lower shortly after the event.
- Imports/GDP: mixed effects.
- In-space placebo dynamic p-values:
  - For GDP, in the first two years after the start of the event, about 40 percent of placebos result in an effect at least as large as the one estimated for Fiji.
  - For volatility, at least 50 percent of placebos result in an effect at least as large.
  - For imports, only in the first year 20 percent of placebos have an effect at least as large as the one obtained for Fiji.
- Interpretation: Given limited number of placebos, results tentatively suggest a short-term positive impact on GDP and no clear impact on GDP volatility and imports to GDP ratio.
- Macroeconomic context:
  - Diversification episode started after political turmoil in early 2000 (a coup) that slowed structural agenda and reforms.
  - One key goal was to restructure the sugar industry; plans established by 2004.
  - Costly public utilities and limited/deteriorated infrastructure constrained growth.
  - Episode might reflect decline in the sugar industry rather than policies promoting other sectors, possibly explaining limited and imprecise growth impact.
- Synthetic Fiji composition (for GDP panel): Guyana (41%), St. Lucia (26%), Jamaica (21%), Cabo Verde (9%), Trinidad and Tobago (2%), Bahrain (1%).
- Synthetic Fiji composition (for volatility panel): Vanuatu (38%), Seychelles (24%), Dominican Republic (20%), Cabo Verde (18%).
- Synthetic Fiji composition (for imports/GDP panel): Cabo Verde (40%), Haiti (42%), The Bahamas (12%), Suriname (4%), Bahrain (2%).
- Volatility metric: rolling 3-year standard deviation of GDP per capita.

### Samoa, exports diversification episode, 2008-2014 — Second diversification episode in a row
- Episodes identified: 2001-2004 and 2008-2014; focus on 2008-2014 due to data limitations.
- Exports diversification index: improved by 14.8 percent, from 3.98 to 3.39.
- GDP impact:
  - Actual Samoa has a slightly higher GDP than synthetic Samoa.
  - Average improvement: 0.2 percent of GDP per one percent improvement in the exports diversification index.
  - Dynamic p-values: lie between 60-80 percent, indicating the measured effect is likely to be random.
- Volatility:
  - Some evidence of lower volatility after the diversification episode, likely significant in the first three years after the event.
- Imports/GDP:
  - Sustained reduction of about 5 percentage points, or on average about 1 percent per one percent improvement in diversification.
  - Likely significant with most p-values in the 20-40 percent neighborhood.
- Macroeconomic context:
  - Late 1990s/early 2000s: structural transformation toward a service-based economy via commerce and tourism.
  - Major growth constraints: use of customary land, finance to the private sector, ease of doing business.
  - End-2009 tsunami: physical damage estimated at around 10 percent of GDP; this natural disaster could dampen estimated benefits from the diversification episode.
- Synthetic Samoa composition (for GDP panel): Marshall Islands (56%), Kiribati (30%), Trinidad and Tobago (8%), Belize (6%).
- Synthetic Samoa composition (for volatility panel): Marshall Islands (68%), Fiji (12%), Grenada (7%), St. Kitts and Nevis (9%), Guinea-Bissau (4%).
- Synthetic Samoa composition (for imports/GDP panel): Papua New Guinea (42%), Trinidad and Tobago (39%), Seychelles (17%), Maldives (2%).
- Volatility metric: rolling 3-year standard deviation of GDP per capita.

### Marshall Islands, domestic diversification episode, 2010-2013
- Episode identification: one domestic diversification spell; domestic diversification improved by:
  - 29.5 percent using the 3-sector index.
  - 8.9 percent using the value-added index.
- SCM setup: 2010 used as beginning of the episode (year identified with the 3-sector index).
- GDP impact:
  - After a small positive effect in the year the episode started, actual Marshall Islands had lower GDP than the estimated counterfactual.
  - Estimated effect likely insignificant; p-values in the 60-90 percent range.
- Output volatility and imports ratio:
  - Estimated effects are volatile (changing from positive to negative) and likely insignificant given high and volatile p-values.
- Macroeconomic context:
  - Economy recovering from the 2008 recession triggered by a commodity-price shock.
  - Political instability and high government turnover hindered reform agenda.
  - Tapering in foreign grants slowed recovery momentum.
  - High import-dependence remained an unaddressed vulnerability.
  - Private sector growth hindered by inadequate infrastructure and lack of access to domestic financing.
  - Policy advice: implement broad-based policies (efficient infrastructure services, strengthen education, improve access to secured commercial lending).
  - During period: commercial fisheries expanded and diversified harvesting, processing, transshipment, and exports.
  - Participation in the Parties to the Nauru Agreement (PNA) considered positive for long-term growth and sustainability of the industry.
- Synthetic Marshall Islands composition (for GDP panel): Micronesia (52%), Comoros (40%), Samoa (8%).
- Synthetic Marshall Islands composition (for volatility panel): Vanuatu (63%), Micronesia (32%), St. Lucia (4%), Belize (1%).
- Synthetic Marshall Islands composition (for imports/GDP panel): Seychelles (40%), Kiribati (30%), Micronesia (30%).
- Volatility metric: rolling 3-year standard deviation of GDP per capita.

### Samoa, domestic specialization episode, 2011-2019
- Episode identification: one domestic specialization episode, 2011-2019.
- Domestic diversification index: increased from 0.29 to 0.34, a worsening in diversification of 20.2 percent.
- GDP impact:
  - Compared with synthetic control, Samoa has lower GDP.
  - Effect is significant the first three years after the start of the episode.
  - Estimated effect bottoms out 3 years after start at about -0.6 percent per percentage worsening in diversification.
- Volatility:
  - Significant reduction estimated during years 3-4 after the start of the event.
  - Effect unlikely to be significant in the rest of the evaluation period.
- Imports ratio: estimated effect is small and insignificant.
- Macroeconomic context:
  - End-2009 tsunami caused long-lasting scarring and slowed growth relative to previous 10 years.
  - Reforming state-owned enterprises prioritized for restoring growth; progress in 2011 with privatization of the telecom company and other institutional reforms.
  - Customary land use: institutional updates between 2009-2010.
  - End-2012 cyclone caused further destruction and reconstruction, potentially concentrating activity in construction and reducing diversification; post-2012 estimates cannot be separated from the effect of this second natural disaster.
- Synthetic Samoa composition (for GDP panel): Sao Tome and Principe (58%), Vanuatu (13%), Micronesia (11%), Palau (8%), Fiji (5%), St. Kitts and Nevis (4%), Grenada (2%).
- Synthetic Samoa composition (for volatility panel): Vanuatu (54%), Fiji (40%), Grenada (7%).
- Synthetic Samoa composition (for imports/GDP panel): Mauritius (35%), Dominican Republic (26%), Trinidad and Tobago (19%), The Bahamas (12%), Marshall Islands (8%).
- Volatility metric: rolling 3-year standard deviation of GDP per capita.

### V. Policy Framework — Designing growth strategies for PICs
- Framework components:
  - Revisit the role of broad-based (horizontal) policies and their potential benefits for PICs.
  - Propose use of the industrial policy conceptual framework (Cherif and others (2022)) adapted to PICs’ characteristics.
  - Sketch application to a sector of interest for PICs.
- Horizontal versus vertical policies:
  - Horizontal policies: broad policies creating enabling environment across sectors (strengthening human capital, improving physical infrastructure and access to finance, improving governance and institutions, reducing trade barriers).
  - Vertical policies: targeted sector-specific interventions (e.g., FDI promotion in a specific sector, production subsidies, public investment in R&D).
- Recommended prioritization:
  - Given slow progress toward SDGs, PICs should continue engaging in horizontal policies with broad-based benefits.
  - Improving access to water and sanitation, electricity, and human capital are key areas likely to increase productivity across the board.
  - Focusing on vertical policies without improving key areas is unlikely to yield high and sustained returns.
- Climate and disaster vulnerability:
  - PICs are highly vulnerable to tropical cyclones, tsunamis, earthquakes, volcano eruptions, and to climate change impacts (rising sea levels, temperatures, changing rain patterns).
  - Reducing vulnerabilities is a medium-term priority to lessen disruptions in human and physical capital accumulation and improve potential growth.
  - Investing in resilient infrastructure early and anticipating challenges would smooth reactions to disasters and minimize disruptions to development.
- Financing SDG and resilience investments:
  - Additional spending needed to meet SDGs and reduce climate-related vulnerabilities will require boosting tax revenues to create fiscal space.
  - Sy and others (2022) outline fiscal reforms that could increase tax revenue by an additional 3 percent of GDP and resolve bottlenecks to revenue collection.
- Role of fundamentals for vertical policies:
  - Vertical policies have stronger effects in countries with better policy fundamentals (Baquie and others 2025).
  - PICs should improve governance and institutional quality before or alongside vertical interventions.
- Roadmap principles for implementing vertical policies (adapted from Cherif and others (2022)):
  1. Identify the externality the vertical policy aims to address: nature, magnitude, and required intervention. Target sectors with higher distortions where potential gains are larger; consider regional initiatives to expand market scale and share fiscal burden.
  2. Ensure interventions pass a cost-benefit analysis: quantify social benefits and costs, alternative uses of resources, potential spillovers, volatility reduction, and distributional implications; perform careful costing exercises.
  3. Account for government failures: anticipate rent-seeking, corruption, mercantilist pressures; implement monitoring across conception, execution, and conclusion; improve institutional quality and capacity to reduce government failure risks.
  4. Balance addressing market failures while mitigating government failure risks: include performance targets and exit strategies, maintain competition, focus on exports potential, and strengthen governance to reduce state capture.
- Policy tools taxonomy reference:
  - Cherif and others (2022) categorize policy tools into main groups: 1. Product market tools; 2. Capital market tools; 3. Labor market tools.

*IMF Working Paper — Annex IV case studies on exports partner diversification.*

### 4. Land market tools; and 5. Technology-related tools. These tools can complement or offset one another, and

### 4. Land market tools; and 5. Technology-related tools. These tools can complement or offset one another, and

### Suitability of tools for PICs
- Some policy instruments are more appropriate for Pacific Islands Countries (PICs) given their unique characteristics; Annex V summarizes available instruments and scope/limitations for PICs.
- Cross-country evidence for each instrument is referenced to Cherif and others (2022).
- Baquie and others (2025) find that the link between vertical policies and competitiveness is stronger for products with higher initial competitiveness, indicating that distance to the technological frontier is key. This implies that understanding the competitiveness of PICs’ export baskets matters for designing vertical policies.
- Constraints for PICs:
  - Limited financing options and narrow tax bases (see Sy and others, 2022).
  - Policy initiatives with a fixed cost and horizon (for example, a credit line for exports) could be more beneficial than broad tax exemptions that erode the tax base.
  - Regional initiatives to promote knowledge spillovers and share fiscal burden have potential where common interest exists (for example, developing the Blue Economy).
  - Innovative domestic financing approaches (e.g., National Provident Funds acting as or funding venture capital firms) can promote domestic start-ups if well-designed to limit risk to the NPF while diversifying their portfolio.
  - Improving SOEs’ mandates and governance can boost their role in domestic production.
  - Examples of novel approaches: Solomon Islands implemented a regulatory sandbox to promote innovation in the financial inclusion space.
  - Existing regional synergies: cooperative management of oceanic fisheries through the Pacific Islands Forum Fisheries Agency and the PNA; Pacific Islands Forum efforts on governance and the 2050 Strategy.

### Implications for land market and technology-related tools
- Land market tools and technology-related tools can be used flexibly to complement or offset one another and can serve multiple purposes beyond their immediate objectives.
- Given fiscal constraints, prioritize instruments with clear fixed costs, horizons, and measurable outcomes; emphasize regional cooperation to spread costs and knowledge.

### Policy design considerations
- Emphasize well-targeted, time-bound interventions with clear exit strategies to avoid eroding scarce fiscal resources.
- Leverage regional institutions (for example, FFA, PNA) to share monitoring, knowledge, and market development responsibilities.
- Use regulatory sandboxes and other enabling regulatory frameworks to foster technology-enabled innovation in priority areas like financial inclusion.

---

### Decision-making framework
- No one-size-fits-all growth strategy for PICs; resources, institutional capacity, and political priorities differ across countries.
- Cherif and others (2022) propose addressing three sets of questions (Table 5): 1. Targeting; 2. Implementation; and 3. Governance.
- The framework’s core questions (as presented in Table 5):
  - Targeting:
    - Which sectors to assist?
    - Identify market failures and social costs.
    - What resources are available?
    - What are the country’s priorities?
  - Implementation:
    - How to assist the sectors?
    - How to implement public support to minimize government failures?
    - How to structure, monitor and evaluate interventions? For how long?
  - Governance:
    - Who decides which sectors to assist?
    - How is informational uncertainty reduced?
    - How do governance structures proposed minimize government failure?

---

### Box 1 — Illustrative application to Fisheries and the Blue Economy

Step 1: Identify Market Failures and Externalities
- The tragedy of the commons: unregulated common resources lead to overfishing and depletion of marine life, endangering sustainability of marine food supply.
- Learning by doing: firms do not internalize benefits from experimentation to identify local value addition or alternative energy programs to reduce imported fuel dependence.
- Learning externalities: developing market relationships for new higher value-added fishing products.
- Technology upgrades: limited access to R&D and new technologies to increase value-added; limited financing for new projects.
- Connectivity: proximity to processing facilities and limited connectivity to fish markets impede value chains.

Step 2: Identify Sector-Specific Policy Failures
- Trade policy: assess high tariffs on critical inputs for fish processing, reassess trade agreements and exchange rate valuation.
- Infrastructure: limited and non-climate-proofed infrastructure to support higher value-added enterprises and connectivity.
- Regulations: reexamine regulations related to fishing practices, subsistence and industrial fisheries; evaluate land allocation regulations for factories and expedite customs procedures.
- Skills: scarce specialized labor in sustainable fisheries management and value-adding techniques.
- Monitoring: lack of data on employment by fishery (some data for tuna), lack of data on sustainability of coastal fisheries harvest levels, difficulty monitoring fishing vessels.

Step 3: Coordinated Action Plan to Tackle Market and Policy Failures
- Promote climate-proofed infrastructure and improve existing infrastructure to encourage industrial fishing operations; possible instruments include tax holidays or credit-guarantees against infrastructure updates.
- Enhance the role of FFA, PNA and TVMA in cooperation, knowledge sharing on sustainable practices, export market development, and vessel monitoring.
- Promote Economic Partnership Agreements (EPA) with potential export markets to develop fish processing operations (as done by PNG and Fiji with the EU); concentrate processing in a few countries with appropriate infrastructure while guaranteeing supply from other countries without scale for canneries.
- Perform regular HIES (household income and expenditure surveys) to evaluate progress and effects on big and small-scale fisheries.
- Create a regional incubator to develop pond aquaculture and diversify coastal fisheries.
- Create mobile training institutes to improve post-harvest methods to extend shelf-life of fish from coastal areas and promote domestic fish trade.
- Improve transparency of fisheries-related subsidies and implement an exit-strategy when adequate.
- Continue implementing horizontal policies that improve infrastructure, access to electricity, governance, and transparency; streamline and simplify regulations.

Step 4: Market Signal Feedback to update the Coordinated Action Plan
- Define evaluation methods before implementation to ensure required data is produced.
- Assess program performance at regular intervals based on objective market signals such as production and exports growth.
- If targets are not met, determine whether bottlenecks or inadequate incentive design are responsible.
- Continuously adapt policy tools to changing local and international conditions.

---

### Conclusion — substantive findings and policy implications
- There is no single growth strategy suitable for all PICs; country-specific factors matter.
- Empirical results indicate:
  - For some PICs, a tourism-led strategy could provide good results for higher growth rates if centered around increasing value-added and the exports share of tourism.
  - For most PICs, tourism-led strategies have limitations due to large investment needs or prior tourism-specialization; diversification-based strategies may offer larger benefits.
  - Quantitative case studies show exports diversification strategies have the most impact on lifting growth and reducing volatility.
  - Little evidence that strategies that diversify only domestic output have a favorable impact on growth, though they may reduce output volatility and deliver benefits in variables not analyzed (for example, food security).
  - Building resilience to climate change is a pre-condition for successful diversification-based growth strategies because damage and reconstruction from natural disasters disrupt diversification efforts.
- Policy recommendations:
  - Continue and prioritize horizontal policies (education, health, infrastructure) to build human and physical capital and resilience to natural disasters—these increase capacity across the board though they do not directly increase diversification.
  - Targeted vertical policies that diversify exports are likely to deliver the most growth impact and should complement horizontal policies.
  - Set up coordinated action plans with clear measurement, evaluation, and exit strategies to avoid past pitfalls of diversification efforts.
  - Promote joint and regional action to overcome market failures linked to geography and small domestic markets.

*IMF Working Paper excerpt: Growth Strategies and Diversification in the Pacific Islands Countries*

### Annex I. Country samples.

### Annex I. Country samples.

### Pacific Islands Countries (PICs) included in the sample
- Fiji
- Kiribati
- Marshall Islands
- Micronesia, Fed. States of
- Nauru
- Palau
- Papua New Guinea
- Samoa
- Solomon Islands
- Tonga
- Tuvalu
- Vanuatu

### Comparator groups and countries (as listed)
- PICs Sub groups: Tourism-based; Commodities exporter
- Caribbean Small States (sample includes 15 countries listed): Antigua and Barbuda; Bahamas, The; Barbados; Belize; Dominica; Dominican Republic (note: “11” appears adjacent in the source table); Grenada; Guyana; Haiti; Jamaica; St. Kitts and Nevis; St. Lucia; St. Vincent and the Grenadines; Suriname; Trinidad and Tobago.
  - Note in source: “While the Dominican Republic is larger in terms of GDP and more diversified than other SIDS, we kept it in the comparator sample since the analysis is performed in per capita terms, for which Dominican Republic is not an outlier in the SIDS group.”
- Other SIDS / Other Asian LICs (listed): Bahrain; Cabo Verde; Comoros; Guinea-Bissau; Maldives; Mauritius; Seychelles; Singapore; São Tomé and Príncipe.
- Other Asian LICs (listed): Bangladesh; Bhutan; Cambodia; Lao P.D.R.; Myanmar; Nepal; Timor-Leste, Dem. Rep. of; Vietnam.

*Italic source: Annex I. Country samples, from the IMF Working Paper "Growth Strategies and Diversification in the Pacific Islands Countries" (Working Paper No. WP/2025/111).*

---

### Annex II. Diversification Episodes

### Method for identifying diversification episodes (algorithm summary)
- Apply an HP filter to the corresponding diversification index, using a smoothing parameter equal to 100 for annual data.
- Identify and estimate the number of unknown breaks using Bai and Perron (1998)’s algorithm, as implemented in Stata by xtbreak (Ditzen and others, 2021).
  - Trimming option is set to 0.15 (inverse of the interstitiary period, minimum number of years between breaks).
  - Only breaks in the constant are assumed; no covariates are used.
- Classify breaks:
  - Upbreak: break that goes to a higher constant.
  - Downbreak: break that goes to a lower constant.
- Define a diversification (specialization) episode:
  - It starts with a downbreak (upbreak).
  - It ends with:
    1. A downbreak (upbreak) of less than 5%; or
    2. An upbreak (downbreak); or
    3. A trough in the HP filtered data; or
    4. The end of the sample.

### Table A.II.1: Diversification episodes identified for PICs (selected entries preserved verbatim)
- Fiji:
  - Exports diversification: 1969-1972
  - Domestic diversification (UN): 1985-1991, 2001-2007
  - Domestic diversification (WDI): 1971-1984*, 1989-1992, 1998-2014*
  - Exports partner diversification: 2006-2014
- Kiribati:
  - Exports diversification: 1979-1981, 1987-1990, 1995-1998*, 2003-2006, 2009-2014*
  - Exports partner diversification: 2010-2015
- Marshall Islands:
  - Exports diversification: 2002-2005*
  - Domestic diversification (WDI): 2009-2011
  - Exports partner diversification: 2010-2013; 2015-2019
- Micronesia:
  - Domestic diversification (WDI): 2003-2006*, 2009-2012, 2017-2019*
- Nauru:
  - Domestic diversification (WDI): 2003-2010*, 2017-2019
- Palau:
  - Exports diversification: 2011-2015*
  - Domestic diversification (UN): 2005-2014*
- Papua New Guinea:
  - Exports diversification: 1969-1970, 1998-2003
  - Domestic diversification (UN): 2009-2012, 2016-2018*
  - Domestic diversification (WDI): 2010-2013*, 2009-2019
- Samoa:
  - Exports diversification: 1969-1972, 1992-1995*, 2001-2004, 2008-2014, 2018-2019*
  - Domestic diversification (UN): 1980-1982, 1992-1995*, 2001-2004, 2008-2014, 2011-2019*
  - Domestic diversification (WDI): 2002-2004, 2014-2018
- Solomon Islands:
  - Exports diversification: 1976-1978, 1992-1995*, 2003-2006*
  - Exports partner diversification: 2012-2015
- Tonga:
  - Exports diversification: 1978-1982, 1991-1995*, 2002-2012, 2015-2019*
  - Domestic diversification (UN): 2002-2009
  - Domestic diversification (WDI): 2000-2008*
- Tuvalu:
  - Exports diversification: 1984-1987, 1998-2002, 2005-2014*
  - Exports partner diversification: 2013-2015, 2018-2019*
- Vanuatu:
  - Exports diversification: 1969-1974*, 1988-1994, 2005-2009*
  - Exports partner diversification: 2016-2018
- Note: * refers to a specialization episode.

*Italic source: Annex II. Diversification Episodes, from the IMF Working Paper "Growth Strategies and Diversification in the Pacific Islands Countries" (Working Paper No. WP/2025/111).*

---

### Annex III. Synthetic Control Method (SCM) — application and goodness-of-fit

### Purpose and application
- SCM used to estimate effects of diversification episodes on:
  - Real GDP per capita (PPP 2011 international dollars, per capita)
  - GDP volatility (rolling 3-year standard deviation)
  - Imports/GDP ratio
- Donor pool: sample of countries in Small Islands Developing States (see Annex I).
- Exclusions from donor pool:
  - Countries that started a diversification episode within a 3-year window of the treated episode.
  - Countries affected by conflict (internal or international).
- Matching targets for pre-treatment fit:
  - Average number of natural disasters in the pre-treatment period.
  - Each of the outcome variable realizations in the pre-treatment period.
- Pre- and post-treatment windows:
  - Use 10 pre-treatment years where possible.
  - Keep number of post-treatment years constant at 5.
  - For some volatility counterfactuals, use the maximum available pre-treatment years if less than 10 but 10 years of pre-treatment data available for GDP effects.

### Inference and fit metrics
- Inference: permutation method (placebo experiments) following Abadie, Diamond and Hainmuller (2010).
- Two fit measures presented:
  - Fit index (closer to 0 indicates better fit).
  - Relative fit index (proportion of placebos with pre-treatment RMSPE at least as large as the treated unit; closer to 1 indicates better fit).
- Adjusted p-values computed to account for pre-treatment fit quality for placebos.

### Table A.III.1 — Goodness of fit and fitted model (selected entries)
- Fiji, 2001 — GDP, 2011 PPP dollars, per capita
  - Pre-RMSPE: 154.4978
  - Fit index: 0.0158
  - Relative fit: 0.6842
  - Model: Average natural disasters. 10 pre-years.
- Fiji, 2001 — GDP volatility, 3-year rolling window
  - Pre-RMSPE: 66.8338
  - Fit index: 0.2587
  - Relative fit: 0.4118
  - Model: Average natural disasters. 9 pre-years.
- Fiji, 2001 — Imports to GDP ratio
  - Pre-RMSPE: 2.0972
  - Fit index: 0.03855
  - Relative fit: 0.8462
  - Model: Average natural disasters. 10 pre-years.
- Samoa, 2008 — GDP, 2011 PPP dollars, per capita
  - Pre-RMSPE: 84.6961
  - Fit index: 0.0177
  - Relative fit: 0.6875
  - Model: Average natural disasters. 10 pre-years.
- Samoa, 2008 — GDP volatility, 3-year rolling window
  - Pre-RMSPE: 48.7015
  - Fit index: 0.3108
  - Relative fit: 0.50
  - Model: Average natural disasters. 8 pre-years.
- Samoa, 2008 — Imports to GDP ratio
  - Pre-RMSPE: 3.4359
  - Fit index: 0.0690
  - Relative fit: 0.5294
  - Model: Average natural disasters. 10 pre-years.
- Marshall Islands, 2010 — GDP, 2011 PPP dollars, per capita
  - Pre-RMSPE: 88.1608
  - Fit index: 0.0276
  - Relative fit: 0.6471
  - Model: Average natural disasters. 10 pre-years.
- Marshall Islands, 2010 — GDP volatility, 3-year rolling window
  - Pre-RMSPE: 47.1609
  - Fit index: 0.4827
  - Relative fit: 0.5789
  - Model: Average natural disasters. 10 pre-years.
- Marshall Islands, 2010 — Imports to GDP ratio
  - Pre-RMSPE: 7.9032
  - Fit index: 0.0833
  - Relative fit: 0.1667
  - Model: Average natural disasters. 10 pre-years.
- Samoa, 2011 — GDP, 2011 PPP dollars, per capita
  - Pre-RMSPE: 72.4981
  - Fit index: 0.014
  - Relative fit: 0.7333
  - Model: Average natural disasters. 10 pre-years.
- Samoa, 2011 — GDP volatility, 3-year rolling window
  - Pre-RMSPE: 43.4618
  - Fit index: 0.2640
  - Relative fit: 0.6875
  - Model: Average natural disasters. 10 pre-years.
- Samoa, 2011 — Imports to GDP ratio
  - Model: Average natural disasters. 10 pre-years. (Pre-RMSPE and fit metrics not listed in source excerpt.)

*Italic source: Annex III. Synthetic Control Method, from the IMF Working Paper "Growth Strategies and Diversification in the Pacific Islands Countries" (Working Paper No. WP/2025/111).*

---

### Annex IV. Exports Partner Diversification Episodes — selected case studies and findings

### Overview (as stated in source)
- Measure: Theil index of exports partner diversification (3-year moving average).
- Interpretation: A lower value of the index represents an increase in diversification.
- Main finding (case-study summary): Trading partners diversification has had some effectiveness in reducing volatility in output, imports, and exports, but limited to no impact in terms of improving GDP.

### Fiji — exports partner diversification, 2006-2014
- Identified episode: 2006-2014.
- Measured improvement: exports partner diversification index improved by 11% from 1.57 to 1.41.
- SCM estimated impacts:
  - GDP: Actual Fiji has a lower GDP than the synthetic one; estimated effect is inconclusive to insignificant.
  - GDP volatility: Lower after the event, with p-values below 20% (source: “p-values below 20%, likely significant for the small sample considered”).
  - Imports ratio and imports volatility: No significant effect.
  - Exports volatility: Very short-run significant reduction, but effect is inconclusive or insignificant after the first year.
- Macroeconomic context (source text):
  - Growth sluggish due to political developments, delays in structural reforms, and a worsening in the terms of trade.
  - Key sectors under pressure: sugar production and garments industry (declines in trade concessions).

- Synthetic Fiji composition in SCM panels (as listed):
  - Panel 1 (GDP): Belize (35%), Jamaica (29%), St. Lucia (29%), St. Kitts and Nevis (6%), Maldives (1%).
  - Panel 3 (GDP volatility): Belize (38%), Jamaica (31%), Guyana (19%), Sao Tome and Principe (7%), Micronesia (3%), Seychelles (2%).
  - Panel 4 (Imports/GDP ratio): Cabo Verde (93%), Antigua and Barbuda (6%), Maldives (1%).
  - Panel 5 (Imports volatility): Barbados (42%), Suriname (33%), Cabo Verde (19%), Jamaica (6%).
  - Panel 6 (Exports volatility): Suriname (41%), Jamaica (31%), Seychelles (28%).
- Volatility definition in panels: rolling 3-year standard deviation of GDP per capita (or of Imports/Exports in US dollars where specified).

### Marshall Islands — exports partner diversification, 2015-2019
- Identified episode: 2015-2019.
- Measured improvement: 23% improvement in diversification.
- SCM estimated impacts:
  - Short-run significant effect on GDP volatility.
  - For GDP level, imports/GDP, and exports/imports volatility: effects either inconclusive or insignificant.
- Macroeconomic context (source text):
  - Heavy dependence on external aid.
  - Sluggish private sector development due to remoteness, dispersion, and regulatory weaknesses.
  - Growth volatile, driven alternatively by fisheries and public administration services.
  - Policy priorities: reform of state-owned enterprises and adaptation to climate change.

- Synthetic Marshall Islands composition in SCM panels (as listed):
  - Panel 1 (GDP): Micronesia (53%), Kiribati (30%), Papua New Guinea (10%), Fiji (4%), Guyana (2%), Palau (1%).
  - Panel 3 (GDP volatility): Kiribati (72%), Tonga (18%), Papua New Guinea (7%), The Bahamas (2%), Mauritius (1%).
  - Panel 4 (Imports/GDP): Kiribati (83%), Micronesia (17%).
  - Panel 5 (Imports volatility): Palau (72%), Micronesia (28%).
  - Panel 6 (Exports volatility): Kiribati (74%), Tonga (15%), Palau (10%), Guyana (1%).
- Volatility definitions: rolling 3-year standard deviation of GDP per capita (or of Imports/Exports where specified).

*Italic source: Annex IV. Exports Partner Diversification Episodes, from the IMF Working Paper "Growth Strategies and Diversification in the Pacific Islands Countries" (Working Paper No. WP/2025/111).*

---

### Annex V. Vertical Policy Tools — scope and limitations for PICs (summary of table)

For each policy tool type, the source lists the tool, a short description, and scope/limitations specific to PICs. Selected entries preserved verbatim.

- Product Market — Trade Policy:
  - Tool description: differential tax rates for profits from exports sales, import-tariff rebates, credit lines for exports, regional trade agreements.
  - Scope and limitations for PICs:
    - Measures to promote exports, encourage participation, or promote import substitution. Must be carefully designed to ensure consistency with World Trade Organization Rules and other trade agreements.
    - Industry case studies find that they often lead to net welfare losses, despite promoting growth.
    - They can limit technology adoption and leapfrogging if tariffs applied on capital and intermediate goods.
    - Can entail a large fiscal cost in a limited tax base environment.
    - Regional agreements can foster exports growth by virtually increasing the scale of the domestic market.

- Product Market — Tax incentives to promote investment:
  - Tool description: tax holidays and exemptions, special corporate tax structures, targeted allowances, subsidized infrastructure.
  - Scope and limitations for PICs:
    - Measures to ease the pressure of a high corporate income tax and attract Foreign Direct Investment.
    - Realizing the full benefits from FDI requires other complementary policies in place (for example, macroeconomic stability, trade openness, stable financial markets, sound business environment).
    - Tax base erosion if not carefully designed and time-bound. PICs already need to rationalize tax exemptions.
    - Not applicable for countries with no or very low corporate income tax.

- Product Market — State as producer and consumer (SOEs):
  - Tool description: State-owned enterprises in key sectors that require large capital investments and fixed costs.
  - Scope and limitations for PICs:
    - SOEs are already in place in key infrastructure sectors.
    - Improving transparency by better governance and clear independent mandates would shield SOEs from misuse.

- Product Market — Measures to reduce informational frictions:
  - Tool description: Initiatives that match buyers and sellers, organize quality certifications etc.
  - Scope and limitations for PICs:
    - One-stop shops are more efficient than multiple agencies.
    - Regional agencies to share exports-related knowledge, build capacity and provide certifications could reduce fiscal costs by sharing the fiscal burden.

- Capital Market — Directed and direct lending:
  - Tool description: Government instructs banks to increase lending to certain sectors or lends directly through development banks.
  - Scope and limitations for PICs:
    - Can boost production if the targeted firms experience severe financial constraints.
    - Potential when development bank already exists and has a clear mandate.
    - Can present fiscal risks if loans are made at subsidized rates.

- Capital Market — Credit guarantees:
  - Tool description: The government provides a loan guarantee to support credit flows.
  - Scope and limitations for PICs:
    - These programs tend to have high fiscal costs and there is reduced evidence of their effectiveness.
    - Creates contingent liability for the public sector.
    - Potential when land reform and collateralized loans are not possible.

- Capital Market — Venture capital and incubators:
  - Tool description: Venture capital provides financing to start-up firms. Incubators provide services (capital, space, expertise).
  - Scope and limitations for PICs:
    - Potential for National Provident Funds to act as Venture Capital firms, when portfolio diversification is needed.

- Labor Market — Skills development:
  - Tool description: Tax credits or subsidies to firms or industry associations that provide training; creation of vocational training geared toward industry-specific skills.
  - Scope and limitations for PICs:
    - Success requires strengthening primary and secondary education.

- Labor Market — Labor taxes:
  - Tool description: Lower labor costs in desired sectors by reducing payroll taxes; provide tax holidays or credits based on employment creation.
  - Scope and limitations for PICs:
    - Carefully designed to satisfy OECD and WTO regulations. Include sunset clauses to limit potential abuses.
    - Not compatible with informal labor markets.

- Land Market — Cheap land:
  - Tool description: Provide access to public land at below market rates.
  - Scope and limitations for PICs:
    - Attractive for foreign investors to bypass negotiation in the local land market.
    - Well-defined land property rights would be preferred.
    - Potential when land tenure reform is hard to achieve.

- Land Market — Special Economic Zones (SEZs):
  - Tool description: Area in which business and trade regulations are different than in the rest of the country.
  - Scope and limitations for PICs:
    - Can provide firms with better infrastructure and public services, and more streamlined regulations.
    - Mixed evidence related to their success.
    - Reduce incentives for comprehensive reforms that could benefit a broader base.

- Technology — R&D incentives:
  - Tool description: Measures to promote R&D, such as tax incentives, subsidies, direct funding, or public-private research centers.
  - Scope and limitations for PICs:
    - Justified by externalities created by new technology.
    - Returns to R&D tend to be smaller in developing countries since they depend on the level of human capital.
    - Technology-transfer instruments can prove more useful for developing countries (government-purchased technology licenses, patent pools, training institutes).
    - Limited scope from direct R&D benefits if not complemented by education and infrastructure improvements.
    - Potential for regional training institutes.

*Italic source: Annex V. Vertical Policy Tools, Scope and Limitations for PICs, from the IMF Working Paper "Growth Strategies and Diversification in the Pacific Islands Countries" (Working Paper No. WP/2025/111).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025111-print-pdf.pdf_