Resilience and Growth in the Small States of the Pacific

Chapter 4. The Pacific Speed of Growth: How Fast Can It Be and What Determines It?

Hoe Khor, Roger Kronenberg, and Patrizia Tumbarello
Published Date:
August 2016
  • ShareShare
Show Summary Details
Yongzheng Yang, Hong Chen, Shiu Raj Singh and Baljeet Singh 

Economic recovery in most Pacific island countries (PICs) following the global financial crisis was mixed.1 Although the region’s two resource-rich countries, Papua New Guinea and the Solomon Islands, rebounded strongly on the back of high commodity prices after the crisis, most other PICs are still struggling with slow growth, with annual GDP growth averaging about 2 percent in 2011–14. Slow income growth has made it more challenging to reduce poverty and youth unemployment, both major socioeconomic issues in the region that need to be urgently addressed.2

With such a tepid recovery, a key challenge is how to raise the growth rate in PICs over the medium term. The slow economic recovery in many of these countries partly reflects their low growth rates when the global financial crisis began—and weak recovery is often associated with weak growth. But what is more worrying is the secular decline of economic growth in the non-commodity-exporting Pacific islands during 2001–14, again, with some exceptions. This is in sharp contrast with developing economies in other parts of the world, most of which managed to raise quite strong growth over this period. Of particular interest is the comparison with low-income countries, which grew three times as fast as PICs during the 2000s after growing at a similar rate during the 1990s.

Does this mean that PIC economies are performing below their potential? A large body of literature examines the economic growth performance of PICs compared with other country groups and relative to their own growth potential. Economic geography suggests that small states such as PICs are disadvantaged in economic growth: their small size makes it harder to exploit economies of scale, and remoteness and insularity increase the cost of transportation for international trade and technology dissemination. The empirical results are inconclusive, however, and complicated by how to take into account the unique characteristics of small states in terms of country size and geographical location.

Against this background, we address the following questions in this chapter: To what extent has economic growth in PICs been slower than in other small states and other similar country groups over the past decades? Have natural conditions, such as small size and geographical location, and external shocks, such as fuel prices and changing trade preferences, played a role in determining the long-term growth of PICs? Why did growth in these countries slow during 2001–10? And how might economic and other policies have affected their growth performance?

We use a unique data set for small states to examine the growth performance of PICs and its determinants in comparison with other countries that have similar characteristics. Previous attempts to analyze growth in PICs tend to lump them together with other small states or examine them in isolation. We also take a long-term perspective in analyzing their growth and pay particular attention to the declining performance during 2001–10. In doing so, we examine the broad international environment as well as internal factors, including through case studies, that may have contributed to the slowing growth.

These questions are put into context by providing an overview of the economic performance of PICs over past decades in a regional and global context, and the growth constraints in small states are briefly reviewed. This is followed by an overview of the methodology and data used in the econometric analysis, and an attempt to identify and quantify growth determinants in small states, and PICs in particular, with the aim of distinguishing the role of natural conditions from policy-related factors. This chapter tries to shed some light on why growth in PICs slowed in 2001–10, and closes with a discussion on the implications of the empirical findings for economic policies.

Growth Performance: Stylized Facts

Our focus on growth performance does not imply that this should be the sole criterion for assessing economic success. As many note, PICs have made considerable progress in many aspects of development, including the Millennium Development Goals. Nevertheless, as recognized in the development literature, inclusive growth is fundamental to poverty reduction and broad human development, and its continued improvement is crucial for sustaining development in the Pacific.

Overall, economic growth in PICs has lagged behind peer groups (Figure 4.1), especially if we exclude Papua New Guinea and the Solomon Islands. In the 1970s PIC economies grew at a respectable rate of just below 4 percent a year, albeit still not as fast as low-income countries (LICs) and other small states on average.3 The 1980s was a slow-growth decade for PICs, but the performance of LICs and small states was also weaker. The 1990s saw PIC growth rebound and nearly catch up with LICs, but this was followed by a sharp divergence during the 2000s. Although the acceleration of LICs contributed most to this divergence, it is also evident that PICs had the weakest growth in four decades, at just one-third the rate of LICs.4 Although growth rates during 2011–14 have been in aggregate higher for the PICs than in the previous decade, the average marks wide differences within the group.

Figure 4.1Annual Average Real GDP Growth


Sources: IMF, International Financial Statistics database; and IMF staff estimates.

Note: ECCU = Eastern Caribbean Currency Union; LICs = low-income countries; PICs = Pacific island countries.

There has been considerable variation in economic growth across PICs (Figure 4.2). The two resource-rich economies in the region, Papua New Guinea and the Solomon Islands, riding world commodity booms and investment in the commodity sector, had a dramatic change in fortune during 2001–14, outperforming their non-resource-rich peers after a long period of weak growth during 1970–90. In fact, Papua New Guinea grew well above the regional average even in the 1990s. Fortunes have also changed in the opposite direction. Fiji, for instance, after outperforming most of its neighbors for most of the 1990s, registered one of the region’s slowest growth rates in the past decade, until recently. As we discuss later, this was largely a result of domestic developments.

Figure 4.2Annual Average Real GDP Growth


Sources: IMF, International Financial Statistics database; and IMF staff estimates.

Note: PICs = Pacific island countries; PNG = Papua New Guinea; SI = Solomon Islands.

PICs’ growth has been volatile but, on average, at similar levels of volatility as other groups of developing economies (Figure 4.3). It is worth noting that growth volatility was high in the 1970s and 1980s, at least in those countries with available data. In the 1990s volatility remained high for all the regional groups except the Eastern Caribbean Currency Union. In the 2000s volatility declined noticeably in all groups except the Eastern Caribbean Currency Union, which experienced its most volatile period in four decades when growth slowed. Growth volatility was lower in PICs in the 2000s, along with the slowest growth rate of the regional groups. Volatility declined in all groups in 2011–14.

Figure 4.3Standard Deviation of GDP Growth

Sources: IMF, International Financial Statistics database; and IMF staff estimates.

Note: ECCU = Eastern Caribbean Currency Union; LICs = low-income countries; PICs = Pacific island countries.

Growth Constraints in Small States

So why have PICs tended to grow less rapidly than peer groups and why did growth slow during 2001–14? In this section, we focus on the first question and leave the discussion of the second to later.

Research on the determinants of growth and volatility in PICs has been closely related to “special problems of small states” (Streeten 1993).5 These are inevitably linked to country size and geographic location and the issues they entail, particularly in PICs. Small populations give rise not only to diseconomies of scale if the domestic market is relied upon as the main source of demand, but also difficulties with industrial clustering, which is used in modern manufacturing (for example, electronics and toys) and service industries (for example, financial and information technology) to reduce production costs.6 The location of many small states, and PICs in particular, makes them remote and insular from major international markets, raising the cost of transportation and communication. Coupled with the effect of country borders (including trade restrictions),7 this means that countries with small populations are less able to exploit economies of scale and hence suffer from lower productivity (if not lower growth of productivity).8 Small states also face high unit costs in providing public goods because of the indivisibility of such goods (World Bank 2010; Commonwealth Secretariat and World Bank 2000; and see Chapter 2 in this volume).

Thus, theory suggests that PICs are particularly disadvantaged by their small populations and location in terms of production costs and economic integration. Figure 4.4 shows that not only are many PICs small, but they also tend to be more remote than most other small countries.

Figure 4.4GDP-Weighted Distance


Source: Gibson and Nero (2007).

Note: PICs = Pacific island countries. Country labels in the figure use International Organization for Standardization abbreviations.

Smallness may not be a disadvantage if a country is surrounded by large, advanced economies (for example, Luxembourg) that offer large markets without significant trade barriers. But in PICs, their small size and location reinforce their disadvantages, and remoteness increases their transportation costs, raising the costs of exporting and importing, which in turn raise the cost of domestic production and exports.9 At the same time, smallness tends to add to transportation costs, either because of weak competition (for example, fewer airlines and shipping companies), higher infrastructure costs (for example, fewer resources for building efficient ports), or small transportation volumes (for example, containers not fully utilized). Distances between the islands of the same country also add to production costs. So, while small can be beautiful, remoteness is not, as far as economic efficiency is concerned.

In contrast to the theory, empirical evidence on the effects of smallness and remoteness on income and economic growth is inconclusive. In an extensive survey, Armstrong and Read (2003) conclude that there is no evidence that small states grow more slowly, despite earlier expectations. Evidence is also weak on the negative impact of “islandness” on growth. This is, however, somewhat surprising given that the literature has demonstrated that landlocked countries—which face similar challenges to island countries in isolation-induced transportation and communication costs—tend to grow more slowly than coastal countries (Gallup, Sachs, and Mellinger 1998; Hausmann 2001). If these conclusions are correct, there must be some intrinsic characteristics of small states that enable them to offset their geography-related disadvantages.

Greater trade openness has been found to be one such offsetting factor. Small states tend to have higher trade-to-GDP ratios than larger countries. Easterly and Kraay (2000) find that the benefits of this greater trade openness offset the adverse impact of greater output volatility in small states.10 In the sample that they used, small states have a ratio of trade to GDP that is 54 percentage points higher than the average economy and the effect of this greater openness is 2½ times larger than the negative effect of greater output volatility. The authors conclude that, even if output volatility is one of the consequences of openness, the greater openness of small states is, on balance, a positive factor for their growth. Of course, greater volatility is not the only (or even the most important) disadvantage that small states face.

Because of greater trade openness, the growth performance of small states is more closely linked to that of their major trading partners. Countries in more dynamic and prosperous regions are likely to grow faster than those in stagnating regions, and vice versa. Empirical research finds supporting evidence for this hypothesis (Armstrong and Read 2000). In a similar vein, Gibson and Nero (2007) find that growth in small states is heavily influenced by the growth of neighboring countries located within 35 degrees of latitude or longitude (a distance of approximately 3,900 kilometers, or 2,423 miles, at the equator).11 Their results show that each percentage point increase in the average growth rate of this neighboring area raises the GDP growth rate of small states in the region by 0.54 percentage point. Looking at this type of growth linkage from a different angle, Bertram (2003) finds that the GDP per capita of small island economies and its growth through time are explained to a large extent by the closeness of the political linkages tying each island to a corresponding advanced economy and the level of GDP per capita in that advanced economy. Estimates show that for every U.S. dollar increase in the GDP per capita in the advanced economy, there is a US$0.30 to US$0.56 increase in the GDP per capita of island states.

Model, Data, and Methodologies

Model Specification

The review on growth constraints in small states suggests that it is not clear how important geography-related disadvantages are in determining the growth performance of small island states relative to common constraints facing all countries. In what follows, we take a more comprehensive approach to the analysis of growth determinants, using a cross-section data set covering 1992–2008 for 45 small states (see Annex 4.1, Table 4.1.1 for the list of countries). Our aim here is to test within this relatively homogeneous group of small island states what differentiates the growth performance of PICs from their peers. Once growth determinants are identified, we evaluate their relative magnitude of impact. This allows us to assess potential benefits from policy actions to influence these determinants. Our approach is to start with parsimonious specifications and move on to more comprehensive and sophisticated specifications.

The model employed in the current study is the growth-initial level model, which is also called the beta-convergence model. The development of the beta-convergence literature has reached a high level of sophistication, owing to the efforts of Mankiw, Romer, and Weil (1992), Barro and Sala-i-Martin (2004), and Islam (2003a, 2003b). The basic form of the growth-initial level equation can be expressed as follows:

where gy is the average growth of GDP per capita over the period under study; the subscript denotes province i; –(1 – eλτ) captures the convergence effect, with –(1 – eλτ) < 0 providing evidence that economies converge to their respective steady-state levels of income, that is, beta-convergence; while –(1 – eλτ) > 0 indicates beta-divergence. The rate of convergence, λ, can be restored from the estimated coefficient, y0 is real GDP per capita in the initial year, s is the share of gross fixed capital formation in GDP, n is the growth rate of the total population, g denotes technological progress, δ is the physical capital depreciation rate, and X is a variable matrix representing variables other than those related to beta-convergence in determining economic growth. In this study we set g + δ = 5 percent.12 Furthermore, by making the restriction that the coefficients of ln(s) and ln(n + g + δ) have identical magnitudes but different signs as in equation 4.1, one is able to obtain the output share of physical capital, α. However, if this restriction is not supported by empirical analysis, as in the current study, the growth-convergence model should have the following expression:

where β1 = –(1 – eλτ).

Data Description

Our data are from the databases of the IMF, United Nations, and World Bank. All money values are in constant 2005 prices in U.S. dollars. In general, 45 small countries are covered in the study, but owing to limitations on data availability for certain series, some regressions cover 39 or 40 countries. Statistics of the relevant series are summarized in Annex Table 4.1.2. The following are used as relevant growth determinants with average real growth of GDP per capita over 1992–2008 (denoted by gy) as the dependent variable:

  • GDP per capita in 1992, denoted by ln(y0), captures the convergence effect.
  • Investment-to-GDP (percent), denoted by ln(investment rate). Values of this variable are averaged over 1992–2008.
  • Population growth together with the physical capital depreciation rate and cost of technological progress (percent), denoted by ln(n + g + δ). Values of this variable are averaged over 1992–2008.
  • The ratio of foreign aid to gross national income (GNI; percent), denoted by Aid/GNI. Because of data limitations, the most recent data (2007) were used. The validity of this proxy is supported by a strong linear correlation between the aid-to-GNI ratio in 2007 and the average aid-to-GNI ratio over different available years during 1992–2007.
  • The political stability series is obtained from the World Bank’s Worldwide Governance Indicators database (Kaufmann, Kraay, and Mastruzzi 2011). The data are averaged over available years during 1996–2008. We also tried other governance indicators from the same source, but only the political stability indicator was found robust.
  • Exports-to-GDP ratio (percent), denoted by Exports/GDP, measures export openness. We also tried imports-to-GDP and trade-to-GDP ratios, but found them less robust. Values of this variable are averaged over 1992–2008.
  • Growth volatility is measured by the standard deviation of GDP growth rates over the period under study. The volatility-growth nexus has been widely discussed in the literature. Some studies, such as Kormendi and Meguire (1985) and Grier and Tullock (1989), found that volatility is positively associated with growth, while others, such as Ramey and Ramey (1995), Aizenman and Marion (1999), and Easterly and Kraay (2000), showed a negative relationship. However, as noted in the literature and our observations in the analysis, the impact of volatility on growth is often linked to some other factors that directly affect growth. For instance, volatility may arise from political instability and trade openness.
  • Remoteness, measured by GDP-weighted distance to the capital cities of major overseas markets, is denoted by ln(distance). We tried several measures of remoteness, such as distance to the top three trading partners; distance to the top three trading partners weighted by GDP; weighted distance to the top three export markets, import sources, and trading partners; average distance to other countries; and distance between producers and consumers within a country. We found only the GDP-weighted distance to the capitals of major overseas markets helped effectively explain growth in the current study.
  • A set of dummy variables to represent continents/regions and the country of Equatorial Guinea, with the latter capturing its exceptional growth with the exploration of oil resources. This set of dummy variables is found to be important in controlling for country heterogeneity and mitigating the heteroscedasticity problem. Apart from this set of dummies, we also tried dummy variables to represent oil producers, oil exporters, commodity exporters, and advanced economies, but none was found significant across various regressions in the current study.

We also considered a number of other factors but found they are statistically insignificant. For example, we tested the importance of smallness as measured by population size. The insignificance of population size in our study of small states, together with the significance of the same indicator in the literature (see, for example, Easterly and Kraay 2000), suggests the impact of smallness on growth may not be linear. This is not surprising given the relatively homogeneous sample used. Other insignificant factors identified include education, inflation, government expenditure, imports, size of agriculture and manufacturing, urban population ratio, arable land ratio, terms of trade, disaster indicators, trade-weighted external demand, and external demand measured by the top three export destinations’ GDP growth. Of course, the performance of these variables could be limited by data quality.

Estimation Methodologies

The analysis in this chapter is based on cross-sectional data on up to 45 countries. Such a relatively small sample can yield unstable results if large heterogeneity exists among observations. However, it is not the case in this study, as our regression results across various specifications are quite consistent, a benefit of the homogeneity of the sample. This benefit is also evident in the consistently high adjusted R-squared values across regressions and the consistency between cross-sectional analysis and panel-data analysis that we carried out, with heterogeneity of countries controlled for by country fixed effects.13

Another issue worth noting is the causality between dependent and explanatory variables. For instance, the causality between investment and growth is believed to be bidirectional. See, for example, Feeny (2005). Endogeneity of aid also receives considerable attention in the aid-growth literature (Mosley 1980; Ali and Isse 2005). Similarly, bidirectional causation between volatility and growth is also widely discussed in the literature (Easterly and Kraay 2000; Malik and Temple 2009). Given this, addressing the endogeneity problem is important in our analysis, as least squares estimates in the presence of endogeneity are biased and inconsistent. Instrumental variables estimators should be applied instead. The Sargan test and Hausman test are employed in our study to detect overidentification of external instruments and endogeneity, respectively. Based on test statistics, we found investment and aid do not pose an endogeneity problem to estimation, while growth volatility does.14,15,16 Therefore, the two-stage least squares estimator is employed to estimate those regressions in which growth volatility is included as a regressor; otherwise the least squares estimator is applied.

The estimation of the specified model follows a simple-to-general strategy; namely, we first test for the neoclassical growth theory by assessing effects of ln(y0), ln(s), and ln(n + g + δ) jointly, and then add and test for additional control factors one at a time.17 This approach helps us identify relevant determinants while avoiding biased estimates caused by omitted relevant variables and minimizing the loss in the efficiency of estimates caused by including irrelevant variables.

Determinants of Growth in Small States

Regression results suggest that geography has a large influence over economic growth in PICs (Table 4.1).18 After controlling for a number of variables found to be statistically significant as determinants of growth, PICs are shown to suffer a distance-related disadvantage in GDP per capita growth of about 1½ percentage points compared with an average non-Pacific small state.19 To put this in perspective, PICs’ annual average GDP per capita growth over 1992–2008 was a little over 0.7 percent. Without the geographical disadvantage, PICs could have grown more than three times as fast in this period.

Table 4.1Regressions on GDP per Capita Growth of Small Island Countries
Estimation Method(1) OLS(2) OLS(3) OLS(4) 2SLS
Variable1Coefficient [t-statistic]Coefficient [t-statistic]Coefficient [t-statistic]Coefficient [t-statistic]
ln(investment rate)1.23[2.11]1.05[1.78]1.08[2.20]1.72[3.20]
ln(n + g + δ)−3.21[-2.04]−4.06[-2.61]−3.94[-3.32]−4.62[-2.84]
Political stability1.25[2.50]1.41[3.19]
Growth volatility−0.36[-2.69]
Sample size45403940
Adjusted/centered R20.81610.87860.90680.8818
Variance inflation factor (mean VIF)2.332.331.99
Breusch-Pagan p-value)0.36340.45340.7104
Sargan statistic0.7387
Source: Authors’ calculations.Note: GNI = gross national income; OLS = ordinary least squares estimator; 2SLS = two-stage least squares estimator. Regressions 1 to 3 employ the ordinary least squares estimator, and regression 4 employs the two-stage least squares estimator.

y0, representing initial per capita income in 1992, measures convergence effect; saving rate is gross fixed capital formation out of GDP; n + g + δ is population growth, g is technological progress rate, and δ is capital depreciation rate. g + δ = 0.05 is adopted in this study. Political stability is one of the governance indicators provided by the World Bank. Growth volatility is measured by the standard deviation of GDP per capita growth rate. Distance is weighted distance to major overseas markets. Dummy variables are included to represent Equatorial Guinea, Pacific island countries, and continents such as Africa, North America, and Europe, which help to control for the heteroscedasticity problem. Growth volatility is found to be endogenous in the current study, and political stability and exports/GDP are used as instrument variables for growth volatility whose validity as efficient instruments is confirmed by the Sargan statistic. Further details on the endogeneity test are in footnote 15. As noted in the regressions, impacts of political stability and the exports-to-GDP ratio not only work on growth directly, but also indirectly by affecting growth volatility. Variance inflation factor (mean VIF) is used to detect the colinearity of the regressors with the constant. A mean VIF of less than 10 can be taken as no evidence of a colinearity problem. The Breusch-Pagan test is employed to test for heteroscedasticity. A p-value greater than a preferred significance level can be taken as no evidence of heteroscedasticity. All regressors are significant at least at the 10 percent level, with most of them significant at either the 1 percent or 5 percent level.

Source: Authors’ calculations.Note: GNI = gross national income; OLS = ordinary least squares estimator; 2SLS = two-stage least squares estimator. Regressions 1 to 3 employ the ordinary least squares estimator, and regression 4 employs the two-stage least squares estimator.

y0, representing initial per capita income in 1992, measures convergence effect; saving rate is gross fixed capital formation out of GDP; n + g + δ is population growth, g is technological progress rate, and δ is capital depreciation rate. g + δ = 0.05 is adopted in this study. Political stability is one of the governance indicators provided by the World Bank. Growth volatility is measured by the standard deviation of GDP per capita growth rate. Distance is weighted distance to major overseas markets. Dummy variables are included to represent Equatorial Guinea, Pacific island countries, and continents such as Africa, North America, and Europe, which help to control for the heteroscedasticity problem. Growth volatility is found to be endogenous in the current study, and political stability and exports/GDP are used as instrument variables for growth volatility whose validity as efficient instruments is confirmed by the Sargan statistic. Further details on the endogeneity test are in footnote 15. As noted in the regressions, impacts of political stability and the exports-to-GDP ratio not only work on growth directly, but also indirectly by affecting growth volatility. Variance inflation factor (mean VIF) is used to detect the colinearity of the regressors with the constant. A mean VIF of less than 10 can be taken as no evidence of a colinearity problem. The Breusch-Pagan test is employed to test for heteroscedasticity. A p-value greater than a preferred significance level can be taken as no evidence of heteroscedasticity. All regressors are significant at least at the 10 percent level, with most of them significant at either the 1 percent or 5 percent level.

As important as it is, geography is not the only factor that has contributed to the slower growth of PIC economies compared to other small states. Growth in small states is influenced by a number of other variables—initial income levels, investment, population growth, aid, export openness, growth volatility, and political stability, all of which were found to be statistically significant. When combined, these variables lower GDP per capita growth in PICs by about another percentage point, compared with an average small state, with some variables making positive contributions and others negative. To start with, PICs’ initial income works to their advantage as they were, on average, poorer than other small states in the early 1990s, the period our initial income calculations is based on. According to the convergence effect found in the regressions, this lower initial income has allowed PICs to grow about ½ percentage point faster than an average non-Pacific small state.

Lower investment (as a percent of GDP) partly explains the slower growth in PICs. 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 been able to achieve the average investment rate of non-Pacific small states, their real GDP growth per capita would have been about ¼ percentage point higher. Given the data limitations, we were not able to disaggregate the investment data into public and private components or by economic sector.20 A detailed analysis is needed to identify at the country level what type of investment would be most productive, but the results here suggest that PICs do have to catch up when benchmarked against other small states. Moreover, there are considerable variations in the investment rate among PICs, which are often as large as between PICs and other small states.

Greater export orientation makes a strong positive contribution to growth in small states. On average, each increase of 10 percentage points in the exports-to-GDP ratio raises GDP growth per capita by about 0.3 percentage point. Since the exports-to-GDP ratio in PICs is 24 percentage points lower than the average of non-Pacific small states, this implies that PICs could have grown by 0.6 percentage point faster had they exported as much as other small states (in percent of GDP). We find that greater imports as a percent of GDP have no statistically significant impact on growth. This suggests that, unlike evidence found for some other countries, imports have not been associated with technological transfers that benefit growth in small states. This, in turn, may reflect the fact that imports in small states more often consist of consumer goods, rather than intermediate inputs or capital goods that embody newer technologies and help improve local productivity. Moreover, imports in smaller PICs are more likely to be linked to aid-funded projects that tend to have weaker linkages with the local economy. Similarly, we also find no evidence that openness measured by the ratio of trade turnover (exports plus imports) to GDP has any positive impact on growth.

Some of the benefits from greater export openness are offset by increased output volatility arising from larger exports (as a percent of GDP). Estimates show that each change of 10 percentage points in export openness is associated with a 0.2 percentage point change in the standard deviation of GDP growth.21 Given this, less than one-fourth of the growth benefits from greater export openness are 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.

Aid is found to be associated with slower GDP growth. In almost all the regressions we tested, the relationship between aid and growth is negative. For each increase of 10 percentage points in aid as a percent of GDP, growth is lower by 0.6 percentage point. Our preliminary tests show that there is no reverse causality. In other words, slower growth does not lead to more aid; it is more aid that leads to slower growth. It is not clear how aid might slow growth in small states, though the most commonly cited mechanism is Dutch disease—a phenomenon of weak export competitiveness resulting from real exchange rate appreciation caused by capital inflows, such as aid. That said, it must be noted that the negative relationship found between aid and growth should not be interpreted as aid lowering economic welfare. In fact, much aid is often aimed at reducing poverty rather than increasing economic growth.22 Unfortunately, we are unable to disaggregate aid into different categories in this analysis, so the estimated impact here applies to total aid rather than to the part of aid that is likely to enhance growth, such as aid used for improving infrastructure and increasing investment in other productive sectors. Furthermore, if aid helps improve living standards in terms of education and health, it could help raise growth in the long term that may not have been captured in this analysis.

Political stability is an important source of faster growth. Measured by the World Bank Political Stability and Absence of Violence/Terrorism indicator, PICs score more favorably on political stability than other small states. This gives PICs, on average, an advantage of 0.3 percentage point in growth over other small states. It should be noted, however, that political stability varies substantially among PICs themselves. Other things being equal, the highest-scoring country in the region has a growth advantage of more than 1 percentage point over the lowest-scoring country.

To sum up the results, PICs have relatively low export openness and investment, but do better than other small states in maintaining political stability. On balance, these policy-related factors, together with geography-related disadvantages, have led to growth rates that are much lower than in other small states. Figure 4.5 shows a rough decomposition of the impact of the various determinants on PICs’ growth, benchmarked against the average growth rate of non-Pacific small states. The first bar depicts the GDP per capita growth rate of non-Pacific small states during 1992–2008. Each of the bars to the right shows, cumulatively, the impact of a growth determinant, with red segments showing the negative impact, blue segments the remaining growth rate after the negative impact, and green segments representing the positive impact. The third bar shows the results when using distance to capture the effect of geography instead of regional dummies, shown in the second bar. The blue bar at the far right shows the actual growth rate in PICs during 1992–2008. The chart reflects that lower export openness, aid, and investment, along with the geographical disadvantage of PICs, are the main contributors to their lower growth compared with other small states.

Figure 4.5Factors Affecting Pacific Islands’ Growth


Source: Authors’ calculations.

Note: See text for explanation of graph.

The results presented in Figure 4.5 should, however, be interpreted with caution. Data quality is always an issue in such exercises, and this is especially true for data on small states. While the results are generally consistent across model specifications, they are subject to data limitations and should be further tested for robustness as data and other information improve. And perhaps even more important, the interpretation of the results should be guided by economic conditions pertaining to the countries in question. For instance, while we found that greater export openness is good for growth, it may not be equally feasible for all PICs to increase exports of goods or nonfactor services, particularly for microstates that have few resources to produce such exports. Similarly, while low investment is generally a constraint on growth in PICs, countries need to identify what impediments investors might be facing and what projects could bring the highest social returns.23 Moreover, any scaling up of investment should take into account debt sustainability if it is financed by borrowing. Capacity constraints at any particular time may also affect the effectiveness of investment.

Why has Growth Slowed in Pics?

The analysis just discussed provides a broad explanation for the relatively slow growth in PICs over the long term, but it sheds little light on why growth slowed in many PICs during 2001–10. In this section, we use the framework established in the previous section to explore possible explanations of the growth slowdown in 2001–08. Statistical tests suggest there were no significant structural breaks in growth for the entire sample of small states. Given this, the same set of regression coefficients from the previous section can be used to predict growth rates based on the level of the determinants in the two subperiods.

It appears that a decline in the exports-to-GDP ratio is a major contributor to the recent growth slowdown. Investment increased slightly, which should have helped raise growth, as should lower output volatility. The positive effects of these developments should add up to over one-half percentage point. On the other hand, the average export openness ratio fell by as much as 4½ percentage points in the 2000s from its 1990s level. There are undoubtedly other factors, including country-specific ones, that have not been included in the model but that contributed to the slower growth in the 2000s. Even so, the decline in export openness, though far from fully explaining the growth slowdown, does point to a weakness in an area that is key to mitigating an obstacle to growth in small states—integration with the global economy.

While the decline in trade openness over 15 years may partly reflect external shocks, it may also indicate weakening competitiveness in some cases. In several PICs, growth began to decelerate well before the increases in world food and fuel prices in 2007–08. At least seven out of the 11 PICs that are the focus of this study had slower growth in 2001–06 than in 1991–2000. For the 11 as a group, average growth in 2001–06 was a little over half the rate of the 1990s and was similar to the average growth of 2007–10, a period full of adverse shocks. In some cases, external shocks seem to have had a lasting impact on productive capacity and weakened growth fundamentals, as in the experience of Samoa following the devastating 2009 earthquake and tsunami and the global financial crisis (Box 4.1). Countries in such circumstances may need to reassess their competitiveness position and adapt their strategies to regain growth momentum.

Real exchange rate appreciation may have played an important part in weakening competitiveness in a number of PICs. In contrast to the 1990s, when only the Solomon Islands experienced an appreciation of its real effective exchange rate, 2001–15 saw the rate appreciate, on average through the entire period, in all six countries with their own currencies, with only Tonga depreciating since 2012 (Figure 4.6).24 With the exception of Papua New Guinea, which has a de jure floating exchange rate regime, the other five countries’ currencies are pegged to a basket of currencies that include the Australian and U.S. dollars, among other key trading partner currencies.25

Figure 4.6Real Effective Exchange Rates of PICs with Central Banks

(Index, January 2000 = 100)

Sources: IMF, International Finance Statistics database; and IMF staff estimates.

Note: PICs = Pacific island countries.

While domestic prices have been rising more rapidly than those in trading partners, there have been limited movements in nominal effective exchange rates in the 2000s, leading to considerable real appreciation in some cases, as in Fiji, Tonga, and Vanuatu, even before the increases in global food and fuel prices. For Papua New Guinea and to a lesser extent the Solomon Islands, the appreciation after 2008 was driven by commodity booms, which boosted export earnings and foreign direct investment inflows. In the five PICs that do not have their own currencies, those that use the Australian dollar (Kiribati, Tuvalu) experienced real appreciation until 2012 (Figure 4.7), whereas those using the U.S. dollar (Marshall Islands, Micronesia, Palau) have seen real appreciation only since 2012 as a result of strengthening of the U.S. dollar.

Figure 4.7Real Effective Exchange Rate of PICs without Central Banks

(Index, January 2000 = 100)

Sources: IMF, International Finance Statistics database; and IMF staff estimates.

Note: PICs = Pacific island countries.

It should be noted that exchange rate appreciation did not in general lead to declines in foreign reserves. In fact, in most cases, reserves as measured by import coverage have risen since the global financial crisis. In some cases, this partly reflects weak import demand resulting from slow economic activity, but also generous donor support.

The evolution of the current account is indicative of the change in external competitiveness over time. While the 2007–08 food and fuel price increases and global financial crisis certainly contributed to this development in PICs in recent years, the deterioration started before these shocks (Figure 4.8). The trade account has also deteriorated (Figure 4.9). It is worth noting, however, that the North Pacific states (Marshall Islands, Micronesia, Palau) saw an improvement in the trade balance in the 2000s. On the other hand, countries that had real effective exchange rate appreciation experienced larger increases in their trade deficits than other countries. Again in the case of Papua New Guinea, it should be noted that the deterioration of the current account also reflects the investment boom associated with its liquid natural gas project. The deterioration of Samoa’s current account was largely due to increases in imports for post-tsunami and more recently post-cyclone reconstruction, although the strong tala may also have helped boost imports. It is also useful to note that the increase in the trade deficit in Fiji, Palau, Samoa, and Vanuatu occurred despite the recent rapid growth of tourism exports.

Figure 4.8Current Account Deficit

(Percent of GDP)

Sources: IMF, World Economic Outlook database; and IMF staff estimates.

Note: PICs = Pacific island countries; PNG = Papua New Guinea; SI = Solomon Islands.

Figure 4.9Goods and Services Deficit

(Percent of GDP)

Sources: IMF, World Economic Outlook database; and IMF staff estimates.

Note: PICs = Pacific island countries; PNG = Papua New Guinea; SI = Solomon Islands.

Box 4.1.Samoa: Managing Shocks and Regaining Its Growth Momentum

The Samoan economy was growing strongly until the mid-2000s when it lost momentum. Well-coordinated reforms in the mid-1990s and sound macroeconomic policies delivered an annual average GDP growth rate of 5 percent during this period. Investment in the lead-up to the 2007 Pacific Games provided another boost before a series of shocks hit the country—the food and fuel price hikes, the global financial crisis, and two devastating natural disasters (a tsunami in 2009 and a cyclone in 2012). Economic growth has since slowed significantly despite a large boost to government expenditure in 2009/10–2012/13, with an average overall fiscal deficit of about 4 percent of GDP. Real GDP growth between 2006/07 and 2012/13 averaged only 0.2 percent, the slowest since the mid-1960s (Figure 4.1.1).

The government’s large infrastructure rebuilding and reconstruction effort following the tsunami and the cyclone was well supported by development partners through grants and concessional loans. However, Samoa’s public debt continued to rise, reaching 55 percent of GDP at the end of 2013/14. Meanwhile, tourism and remittances, two pillars of the economy, have recovered only slowly, and production at the Yazaki automotive component plant, one of Toyota Australia’s biggest component suppliers, has been declining since the global financial crisis. Agricultural production has also been declining for a number of years, though it has seen a recent modest recovery.

Figure 4.1.1Samoa: Actual and Trend Real GDP Growth


Sources: IMF, World Economic Outlook database; and IMF staff calculations.

Samoa now faces a difficult path to wind down its fiscal deficits because the private sector has not picked up the slack. The government is committed to reducing public debt to a more sustainable level to maintain hard-won macroeconomic stability. While the fiscal consolidation is essential for long-term sustainability, it is also important to safeguard economic growth and social spending. Given weak global demand, a great effort will be needed to strengthen competitiveness and revitalize private-sector-led growth.

Samoa’s experience highlights the vulnerability of a small island economy’s competitiveness to exogenous shocks, including natural disasters. Competitiveness can be undermined quickly by such shocks, and it will take coordinated efforts to regain it. A reassessment of the economy’s underlying strength and growth potential is needed to help establish a macroeconomic framework to regain competitiveness and to 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 has a strong track record of reform, however, and the government’s continued commitment to keeping this record bodes well for future progress.

Given the real exchange rate appreciation and deteriorating export performance in many PICs, it is tempting to link weakening competitiveness to the aid-induced Dutch disease effect referred to earlier in the econometric analysis. However, the competitiveness of PICs may have also been affected by the changing external trade environment. As part of the African, Caribbean, and Pacific Group of States, PICs have benefited from nonreciprocal trade preferences under successive Lomé Conventions (first signed in 1975) and subsequently the Cotonou Agreement (signed in 2000). Fiji’s sugar exports, at higher than world prices, are one example of such benefits. But trade preferences under the Cotonou Agreement have been eroded or are being phased out, reducing the export competitiveness of PICs. Nonreciprocal trade preferences offered by Australia and New Zealand under the South Pacific Regional Trade and Economic Co-operation Agreement, which took effect in 1981, are also being eroded as the two countries liberalize their trade. Under the agreement, for example, Fiji was able to export large volumes of garments to Australia and New Zealand, but as preference margins shrank so have Fiji’s exports, especially when restrictions under the Agreement on Textiles, Clothing, and Footwear were phased out in 2005.26 Negotiations on a Pacific Agreement on Closer Economic Relations Plus (PACER-Plus) to achieve the long-term goal of a single regional market among Pacific Islands Forum members have moved slowly.27 In recent years, PICs have not joined any regional trade arrangements in Asia and the Pacific. Among them, only Papua New Guinea is a member of the Asia-Pacific Economic Cooperation (APEC), and none are members of any other major regional grouping or regional trade arrangements in Asia.

The trade patterns of PICs in recent years may have reflected these developments in the trade policy environment. As is the case with all other groups of developing economies, PICs’ traditional markets—the European Union and North America—have become less dominant.28 Developments in the Australian and New Zealand markets have affected resource-rich and non-resource-rich PICs differently. Both markets have become more important for Papua New Guinea and the Solomon Islands, while their importance to other PICs has fallen. In the nontraditional markets, the two PIC groups have also gone in different directions. While Papua New Guinea and the Solomon Islands have shifted to China from other Asian countries,29 the other PICs30 moved in the opposite direction, reflecting the growing importance of southeast Asia’s markets. Non-resource-rich PICs are the only group of countries that have seen less of their exports going to fast-growing China due mainly to their lack of mineral resources and supply constraints on other commodities. Taking Asia as a whole, however, these PICs have diversified their exports to the region.

To fully understand the export performance of PICs, one also needs to take into account service exports, especially tourism. Of the 11 PICs covered in this chapter, five (Fiji, Palau, Samoa, Tonga, Vanuatu) have significant tourism exports, with earnings from the sector accounting for between 10 percent (Tonga) and 70 percent (Palau) of GDP and serving as an important source of economic growth in all five countries.31 Growth as measured by tourist arrivals has varied considerably in this group (Figure 4.10). As with merchandise exports, North America, Europe, Australia, and New Zealand are the traditional tourism markets and still dominate in all countries in the group except Palau, which has diversified into Asian markets in recent years. As Box 4.2 illustrates, both economic policy and geography played an important role in the recent recovery of tourism in the five countries.

Figure 4.10Annual Visitor Arrivals


Sources: Country authorities; and national tourism statistics.

Developments in the cost of international transport may have worked to the disadvantage of PICs in recent years. It is commonly assumed that advances in transport technology should reduce disadvantages facing more remote countries. The introduction of jet engines in the late 1950s and containerized shipping in the late 1950s substantially increased the efficiency of air and ocean transportation. But according to Hummels (2007), technological advances in ocean shipping have been largely trumped by fuel price increases, leaving ocean transport costs (as a percent of the values of shipped goods) in the early 2000s much as they were in the 1950s.32 Moreover, rapid increases in transport costs during 2001–10, especially for air transport, led to a sharp decline in the proportion of airlifted goods in trade after a long period of steady increases (Figures 4.11 and 4.12). For PICs, the increase in air transport costs has added substantially to the transport cost of perishable products. These trends were somewhat mitigated by the substantial fall in global fuel prices starting in mid-2014.

Figure 4.11International Containerized Freight

(Simple average cost in U.S. dollars, TEU deflated by U.S. GDP deflator)

Source: Containerisation International.

Note: TEU = twenty-foot-equivalent unit.

Figure 4.12Trends in Airline Costs

(Index, left scale; percent, right scale)

Other key developments in international transport have also had mixed impacts on trading costs in PICs. It is worth noting that containerized shipping was introduced in PICs much later (mostly in the late 1970s and 1980s) than in other parts of the world because of the substantial investment involved, including in port facilities. Furthermore, competition in the air and ocean transport industries in PICs is limited because of small market size and state monopolies. As a result, the benefits of technological advances may not have been passed on to consumers as much as in other parts of the world.33

Box 4.2.Why the Recovery in Tourism Varies So Much across PICs

Recent experience with the recovery of tourism in PICs shows the importance of linking to Asian markets and the role of economic policy in determining tourism growth. Figure 4.2.1 shows the tourism market by visitor arrivals in Fiji, Palau, Samoa, and Vanuatu.

Figure 4.2.1Tourism Market by Visitor Arrivals

Sources: Samoa Bureau of Statistics; Palau Visitor Authority; Samoa Bureau of Statistics; and Vanuatu National Statistics Office.

Being much closer to Asia than other PICs, Palau has benefited immensely from the region, with Taiwan Province of China and Japan accounting for some two-thirds of the country’s total tourist arrivals in recent years. Tourist arrivals from these and other Asian countries have increased rapidly, helped by more charter flights. However, in 2013, Palau’s tourism industry was hit by the cessation of direct flights from Asia’s emerging market economies, the stronger U.S. dollar, and high tourism-related fees. The industry rebounded in 2014 on strong tourist arrivals from China.

In contrast, Fiji, Samoa, Tonga, and Vanuatu rely on Australia and New Zealand for tourism, which together account for two-thirds to three-fourths of total arrivals to these countries. Fiji’s tourism has grown strongly while that of Samoa, Tonga, and Vanuatu has been slow or stagnating, despite all of these countries benefiting from the sustained growth of Australia and New Zealand. Successful elections in Fiji in September 2014 and its return to democratic government created a favorable environment to attract tourism. The devastating Cyclone Pam in March 2015 is expected to severely impact Vanuatu’s tourism sector in the near term.

As with merchandise trade, exchange rate policies seem to have played an important part in explaining the varying performances of the tourism industries of these countries. The Fijian dollar was devalued by 20 percent against the U.S. dollar in April 2009, and the increased competitiveness from the devaluation temporarily helped boost Fiji’s tourist arrivals, by 17 percent in 2010 and 7 percent in 2011. Palau uses the U.S. dollar as its legal tender, and the currency’s strength since the second half of 2014 relative to Asian currencies has affected Palau’s competitiveness. In contrast, Fiji, Samoa, Tonga, and Vanuatu peg their currencies to a basket of currencies that includes the Australian dollar. The weak Australian and New Zealand dollars since mid-2014 could undermine tourist flows to the Pacific islands from these countries.

There may be substitution among PICs as tourist destinations. This means that if every PIC devalues its currency, the region as a whole may not attract more tourists. However, given the region’s relatively small market share in total Australian and New Zealand tourists, the substitution effect could be easily overstated. In any case, the effect should be limited among Asian tourists.

One of the major benefits from technological advances over the past decades has been faster speed and greater reliability in shipping. This, in principle, would have helped island countries lower transport costs relative to their competitors. However, remote small states such as PICs do not appear to have been able to take much advantage of large, more efficient vessels because of their small trade volumes and the need for ships to stop more frequently to serve small and often widely dispersed destinations. Paradoxically, the technological advances appear to have given a greater advantage to exporters of manufactured goods, as large vessels and greater speed and reliability allow finer segmentation of value chains. Associated with this, the cost advantage enjoyed by high-end goods is growing over time as the spread between high-priced and low-priced goods in each product category widens (Hummels 2007, 2009). Thus, without high-value-added exports, PICs could be increasingly disadvantaged by transportation costs.

Policy Implications

One of the key findings of this econometric analysis is that PICs do seem to face lower speed limits in economic growth because of their remoteness. Moreover, growth volatility in PICs has often been higher than in larger economies. It is important for policymakers to recognize the existence of such limits because unrealistic expectations for growth could lead to overly ambitious targets. Moreover, growth volatility also entails prudent planning and maintaining larger policy buffers in good times.

Admittedly, there are significant challenges in raising growth, including finding a more effective way to allow aid to contribute to long-term domestic productive capacity as well as reducing poverty and fostering inclusive growth. When a country is small and relatively poor, it is likely to receive more aid (relative to GDP) than a large country, and so there is no need to have a more depreciated real exchange rate to keep a balanced trade or current account. This allows small countries to enjoy a higher living standard than they would otherwise, through better public services and cheaper imports. However, it also means that the export sector has to be very efficient to be able to compete on the world market even when it enjoys a comparative advantage. While it seems difficult to fully overcome the effects of aid-supported strong exchange rates, PICs should nevertheless take measures to alleviate them by directing more aid to productivity improvements and accelerating structural reforms.

This analysis highlights the importance and difficulty of macroeconomic policies in creating a competitive environment in small states. As shown, neither a fixed nor flexible exchange rate regime guarantees such an outcome. Most PICs maintain a pegged exchange rate regime, which provides a useful nominal anchor. Historically, however, PICs tend to have higher inflation than their trading partners (as many developing economies do), and periodically the fixed exchange rates become unsustainable and large-step devaluations become necessary to correct external imbalances. On the other hand, a flexible exchange rate (or using a foreign currency) can also lead to rapid appreciation when there are large foreign exchange inflows, such as aid and resource rents. Such appreciation may be necessary for maintaining macroeconomic stability (that is, controlling inflation), but it may not be consistent with the objective of maintaining competitiveness. In principle, such inflows can be sterilized through reserve accumulation, but countries often do not fully sterilize inflows because of the costs involved in mopping up local currency liquidity and the resulting higher domestic interest rates. This tends to cause higher inflation and real exchange rate appreciation. Furthermore, foreign exchange markets are shallow in PICs, and a floating exchange rate regime may result in heavy exchange rate volatility, which would be harmful for trade.

This calls for better coordination of macroeconomic policies to maintain competitiveness and stability. When a country maintains a fixed exchange rate regime, monetary policy must be subordinate to the exchange rate policy and ensure that inflation is not consistently higher than levels in trading partner countries. Similarly, fiscal policy must be prudent to avoid pressure on domestic prices and debt sustainability.34 Large windfall inflows could be saved—for example, through a sovereign wealth fund—for intergenerational distribution and consumption smoothing. Since the global financial crisis, macroeconomic policies in PICs have been geared toward supporting growth through accommodative monetary and fiscal policies. These are appropriate given the circumstances. However, as noted earlier, higher inflation resulting from rising world food and fuel prices has led to significant real exchange rate appreciation in a number of countries since 2008. At the same time, increased aid inflows in response to the global financial crisis have allowed many PICs to maintain or accumulate foreign reserves. However, as economic recovery strengthens with import demand increasing and aid inflows reverting to normal levels, these countries may face challenges to return their real exchange rates to more sustainable and competitive levels.

Structural reforms are critical to mitigating the effects of strong exchange rates through higher productivity. When a country has a fixed exchange rate and faces persistently higher inflation than its trading partners, raising productivity at a pace faster than its trading partners is essentially the only way to avoid periodic devaluations.35 Empirical research indicates that productivity growth in PICs has generally been slow and low investment is a key constraint on higher growth, as demonstrated in the case of Fiji.36 This is consistent with our findings that PICs need to raise the level of investment to grow faster, a self-evident conclusion, but nevertheless a particularly pertinent one given the low investment rates of PICs compared with peer countries. Our findings that political stability is important for growth in PICs are similarly conventional, but they should nevertheless reinforce the resolve of these countries’ governments for better governance.

Improving competitiveness will be difficult without aggressively reducing the costs of distance and insularity, a principal source of the growth disadvantage in PICs. As in the manufacturing sector, the transportation and communication industries also exhibit economies of scale, and this again puts small and remote countries at a disadvantage. The challenge is not only to scale up investment in transportation, communication infrastructure, and connectivity, but also to continue with reforms and appropriate regulation to ensure rigorous competition in these industries. Greater regional cooperation could also help mitigate the effect of diseconomies of scale, as shown by the establishment of the Pacific Forum Line37 and the Pacific Islands Telecommunications Association.38

The ultimate way to overcome smallness and distance is further integration with the global economy, and trade policy can play an important role in this process. PICs need to adapt to a rapidly changing landscape in world trade and finance. As trade preferences under colonial ties are phased out or are being eroded, PICs should seek deeper integration with metropolitan countries, particularly in trade in labor services. Much progress has been made in the temporary migrant workers programs with New Zealand and Australia, but the potential remains large relative to the size of labor forces in PICs. In some, particularly the smaller ones that have fewer natural resources or are constrained in developing such resources, increased trade in labor services provides a critical source of income generation, which in turn boosts domestic activities. Given the limited job opportunities at home, there is little risk of brain drain. PACER-Plus provides a useful framework for further integration with Australia and New Zealand and, to this end, greater efforts could be made to accelerate the negotiations.

Further integration with the global economy should also include strengthening trade and financial ties with Asia. Our analysis suggests that PICs without mineral resources still have untapped opportunity to leverage the fast-growing Asian markets. This is not surprising given that exporting nonmineral products to Asian markets needs reliable supply and marketing. Some emerging market economies in Asia offer duty-free and quota-free entry of goods from the PICs’ least-developed economies.39 However, constraints on domestic supply have meant that these opportunities have not been fully utilized. For the other PICs, it is important to engage with Asian countries to ensure that goods from the Pacific are not discriminated against as Asian countries expand their free trade agreements. Given the supply constraints in PICs, it may seem irrelevant to secure market access to Asia at this stage, but doing so may encourage foreign investment and increase the awareness of goods that PICs can offer. Entrepreneurs in Asia could help bring capital as well as fill in skills and marketing gaps in PICs. As such, a proactive policy toward Asia is important.


In many respects, the findings presented in this chapter are unsurprising, yet they have received little empirical support. PICs do face special challenges in economic growth because of their disadvantages arising from geography. It is also unsurprising that there is plenty of room to raise growth by increasing investment and promoting trade openness as well as ensuring political stability. By reaching the levels of other small island states in these areas, PICs could substantially speed up their economic growth. The key question is how to increase investment and openness. The good news from this study is that lower investment and exports in the Pacific do not seem to be entirely the result of geography. This implies that economic and other policies can make a difference in speeding up growth in the region, including by enhancing connectivity through better composition of public spending. The large variations in the performance across countries provide some evidence that policies do matter. Thus, the main challenge is to identify and implement policies that would be most effective in increasing openness and productive investment based on individual country circumstances.

It is important to recognize the differences in natural endowments and economic circumstances across PICs. Microstates may well face quite different challenges from their relatively large neighbors, principally Melanesian countries. These may necessitate different priorities and approaches to promoting investment and trade. To this end, detailed case studies would be highly desirable for formulating policy recommendations. The fact that growth in most PICs during 2001–14 has lagged behind other small states suggests some fundamental changes may have been taking place—either in policy settings or the external environment—making country size and location more determinant factors than previously thought.

It is tempting to suggest that a competitive exchange rate is essential to inducing higher investment and greater trade openness. While external shocks to trade and commodity prices seem to have contributed to the weakening competitiveness in some PICs, evidence also points to the role played by macroeconomic policies. In the face of substantial size- and geography-induced cost disadvantages and large aid inflows, PICs need to manage their macroeconomic policies to ensure that their exchange rates allow domestic producers to compete in the world market in areas in which they have a comparative advantage. This may prove critical for pursuing private-sector-led growth.

Maintaining competitiveness also entails structural reforms to improve efficiency in production, transportation, and communication. Many reforms have been carried out over the years, and the payoffs have been significant in some areas, as shown by marked declines in the cost of telecommunication following the industry’s deregulation. However, as the quest for faster and more inclusive growth continues, PIC governments will need to keep searching for such high-payoff reforms. The ever changing global environment should only strengthen the resolve of PIC governments to accelerate such reforms.

Annex 4.1. Country List and Data Description
Annex Table 4.1.1Country List
Africa (14)Americas (12)Europe (4)Pacific (9)Asia (6)
BotswanaAntigua and BarbudaEstoniaFijiBahrain
Cabo VerdeThe BahamasMaltaKiribatiBhutan
ComorosBarbadosMontenegroMicronesiaBrunei Darussalam
DjiboutiBelizeSan MarinoPalauMaldives
Equatorial GuineaBermudaPapua New GuineaQatar
The GambiaGrenadaSolomon Islands
LesothoSt. Kitts and NevisVanuatu
MauritiusSt. Lucia
NamibiaSt. Vincent and the Grenadines
São Tomé and PríncipeSuriname
Source: Authors’ compilation.
Source: Authors’ compilation.
Annex Table 4.1.2Average Values of Relevant Series, 1992–2008
SmallNon-Pacific SmallPacific Island
Convergence effect (lnY0)7.988.157.40
Investment [ln(investment/GDP)]3.263.333.04
Population growth [ln(n + g + d)]1.881.881.87
Aid (aid/gross national income, percent)7.965.6920.01
Political stability0.490.530.57
Exports (percent of GDP)52.1156.9735.08
Volatility (growth rate standard deviation, percent)4.614.365.48
GDP growth per capita (percent)2.382.870.64
GDP-weighted distance to major overseas markets (ln)9.018.939.25
Source: Authors’ calculations.
Source: Authors’ calculations.

    AizenmanJoshua and NancyMarion. 1999. “Volatility and Investment: Interpreting Evidence from Developing Countries.Economica66 (262): 15779.

    • Search Google Scholar
    • Export Citation

    AliAbdiweli M. and Hodan S.Isse. 2005. “An Empirical Analysis of the Effect of Aid on Growth.International Advances in Economic Research11 (1): 111.

    • Search Google Scholar
    • Export Citation

    AndersonJames and Eric vanWincoop. 2004. “Trade Costs.Journal of Economic Literature42 (3): 691751.

    ArmstrongHarvey W. and RobertRead. 1998. “Trade and Growth in Small States: The Impact of Global Trade Liberalization.The World Economy21 (4): 56385.

    • Search Google Scholar
    • Export Citation

    ArmstrongHarvey W. and RobertRead. 2000Comparing the Economic Performance of Dependent Territories and Sovereign Microstates.Economic Development and Cultural Change48 (2): 285306.

    • Search Google Scholar
    • Export Citation

    ArmstrongHarvey W. and RobertRead. 2003. “The Determinants of Economic Growth in Small States.The Round Table92 (368): 99124.

    • Search Google Scholar
    • Export Citation

    BarroRobert J. and Xavier I.Sala-i-Martin. 2004. Economic Growth. Second edition. Cambridge, Massachusetts: MIT Press.

    BertramGeoffrey. 2004. “On the Convergence of Small Island Economies with Their Metropolitan Patrons.World Development32 (2): 34364.

    • Search Google Scholar
    • Export Citation

    BowmanChakriya and SatishChand. 2008. “Size Matters: The Impact of Aid on Institutions.Pacific Economic Bulletin23 (2): 14565.

    • Search Google Scholar
    • Export Citation

    ChenHong. 2012. Convergence Total Factor Productivity and Industrial Growth: Growth Empirics in Post-Reform China.Saarbrücken: Lambert Academic Publishing.

    • Search Google Scholar
    • Export Citation

    ChenHong and BaljeetSingh. 2011. “Total Factor Productivity in Fiji over 1983–2007.” In Fiji Economy UpdateSuva, Fiji.

    Commonwealth Secretariat and World Bank. 2000. “Small States: Meeting Challenges in the Global Economy.” From the final communiqué of the Sixty-First Meeting of the Development CommitteeWashingtonApril17.

    • Search Google Scholar
    • Export Citation

    DuncanRon and HaruoNakagawa. 2006. “Obstacles to Economic Growth in Six Pacific Island Countries.Working PaperWorld BankWashington.

    • Search Google Scholar
    • Export Citation

    EasterlyWilliam and AartKraay. 2000. “Small States, Small Problems? Income, Growth and Volatility in Small States.World Development28 (11): 201327.

    • Search Google Scholar
    • Export Citation

    FavaroEdgardoNaomiHalewood and Carlo M.Rossotto. 2008. “From Monopoly to Competition: Reform of Samoa’s Telecommunications Sector.” In Small States Smart Solutions: Improving Connectivity and Increasing the Effectiveness of Public Services edited by EdgardoFavaro. Washington: World Bank.

    • Search Google Scholar
    • Export Citation

    FeenySimon. 2005. “The Impact of Foreign Aid on Economic Growth in Papua New Guinea.Journal of Development Studies41 (6): 10921117.

    • Search Google Scholar
    • Export Citation

    GallupJohn L.Jeffrey D.Sachs and Andrew D.Mellinger. 1998. “Geography and Economic Development.Working Paper 6849National Bureau of Economic ResearchCambridge, Massachusetts.

    • Search Google Scholar
    • Export Citation

    GibsonJohn and Karen L.Nero. 2007. “Are Pacific Island Economies Growth Failures? Geo-Political Assessments and Perspectives.Report for the Pasifika ProjectInstitute of Policy Studies, Victoria University of Wellington.

    • Search Google Scholar
    • Export Citation

    GrierKevin B. and GordonTullock. 1989. “An Empirical Analysis of Cross-National Economic Growth, 1951–80.Journal of Monetary Economics24 (2): 25976.

    • Search Google Scholar
    • Export Citation

    HausmannRicardo. 2001. “Prisoners of Geography.Foreign Policy122 (1): 4453.

    HughesHelen. 2003. “Aid Has Failed the Pacific.Issue Analysis No. 33Centre for Independent StudiesSt Leonards, Australia.

    HummelsDavid. 2007. “Transportation Costs and International Trade in the Second Era of Globalization.Journal of Economic Perspectives21 (3): 13154.

    • Search Google Scholar
    • Export Citation

    HummelsDavid. 2009. “Globalization and Freight Transport Costs in Maritime Shipping and Aviation.Working Paper 3International Transport ForumParis.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2011. “New Growth Drivers for Low-Income Countries: The Role of BRICs.Washington.

    IslamNazrul. 1995. “Growth Empirics: A Panel Data Approach.Quarterly Journal of Economics110 (4): 112770.

    IslamNazrul. 2003a. “What Have We Learnt From the Convergence Debate?Journal of Economic Surveys17 (3): 30962.

    IslamNazrul. 2003b. “Productivity Dynamics in a Large Sample of Countries: A Panel Study.Review of Income and Wealth49 (2): 24772.

    • Search Google Scholar
    • Export Citation

    KaufmannDanielAartKraay and MassimoMastruzzi. 2011. “Worldwide Governance Indicators 2011 Update,WashingtonWorld Bank.

    KleinertJörn and JuliaSpies. 2011. “Endogenous Transport Costs in International Trade.Discussion Papers 74Institut für Angewandte WirtschaftsforschungTübingen.

    • Search Google Scholar
    • Export Citation

    KormendiRoger and PhilipMeguire. 1985. “Macroeconomic Determinants of Growth: Cross-Country Evidence.Journal of Monetary Economics16 (2): 14163.

    • Search Google Scholar
    • Export Citation

    MalikAdeel and Jonathan R. W.Temple. 2009. “The Geography of Output Volatility.Journal of Development Economics90 (2): 16378.

    • Search Google Scholar
    • Export Citation

    MankiwN. GregoryDavidRomer and David N.Weil. 1992. “A Contribution to the Empirics of Economic Growth.Quarterly Journal of Economics107 (2): 40737.

    • Search Google Scholar
    • Export Citation

    McCallumJohn. 1995. “National Borders Matter: Canada-U.S. Regional Trade Patterns.American Economic Review85 (3): 61523.

    MilnerChris and TonyWestaway. 1993. “Country Size and the Medium-Term Growth Process: Some Cross-Country Evidence.World Development21 (2): 20311.

    • Search Google Scholar
    • Export Citation

    MosleyPaul. 1980. “Aid, Savings, and Growth Revisited.Oxford Bulletin of Economics and Statistics42 (2): 7995.

    NobleCameronNataliePereira and NaniseSaune. 2011. Urban Youth in the Pacific: Increasing Resilience and Reducing Risk for Involvement in Crime and Violence.Suva, Fiji: UNDP Pacific Centre.

    • Search Google Scholar
    • Export Citation

    PavlovVlad and CraigSugden. 2006. “Aid and Growth in the Pacific Islands.Asian-Pacific Economic Literature20 (2): 3855.

    RameyGarey and Valerie A.Ramey. 1995. “Cross-Country Evidence on the Link between Volatility and Growth.American Economic Review85 (5): 113851.

    • Search Google Scholar
    • Export Citation

    RaoB. BhaskaraKanhaiya L.Sharma and RupSingh. 2008. “Estimating Aid-Growth Equations: The Case of Pacific Island Countries.Pacific Economic Bulletin23 (2): 8296.

    • Search Google Scholar
    • Export Citation

    ReddingStephen and Anthony J.Venables. 2004. “Economic Geography and International Inequality.Journal of International Economics62 (1): 5382.

    • Search Google Scholar
    • Export Citation

    SrinivasanThirukodikaval N.1986. “The Costs and Benefits of Being a Small Remote Island Landlocked or Ministate Economy.World Bank Research Observer1 (2): 20518.

    • Search Google Scholar
    • Export Citation

    StreetenPaul. 1993. “The Special Problems of Small Countries.World Development21 (2): 197202.

    ThackerNita and SebastianAcevedo. 2011. “A Cross-Country Perspective on Growth in the Caribbean: The Role of Tourism and Debt.Paper presented at the University of the West Indies Central Bank of Barbados and IMF conference“Options for the Caribbean after the Global Financial Crisis,”Bridgetown, BarbadosJanuary 27–28.

    • Search Google Scholar
    • Export Citation

    WeiShang-Jin. 1996. “Intra-National Versus International Trade: How Stubborn are Nations in Global Integration?Working Paper 5531National Bureau of Economic ResearchCambridge, Massachusetts.

    • Search Google Scholar
    • Export Citation

    WintersL. Alan and WonhyukLim. 2010. “Economic Growth in Low Income Countries: How the G20 Can Help to Raise and Sustain It.Economics Department Working Paper Series 8–2010University of SussexBrighton.

    • Search Google Scholar
    • Export Citation

    WintersL. Alan and Pedro M. G.Martins. 2004. “When Comparative Advantage Is Not Enough: Business Costs in Small Remote Economies.World Trade Review3 (3): 34783.

    • Search Google Scholar
    • Export Citation

    World Bank. 2010. “Solomon Islands Growth Prospects: Constraints and Policy Priorities.Discussion NoteWorld BankWashington.

    YangYongzhengHongChenShiu RajSingh and BaljeetSingh. 2013. “The Pacific Speed of Growth: How Fast Can It Be and What Determines It?IMF Working Paper 13/104International Monetary FundWashington.

    • Search Google Scholar
    • Export Citation

    YaoShujie and ZongyiZhang. 2001. “On Regional Inequality and Diverging Clubs: A Case Study of Contemporary China.Journal of Comparative Economics29 (3): 46684.

    • Search Google Scholar
    • Export Citation

This chapter is based on IMF Working Paper 13/104 (Yang and others 2013) prepared for the Joint IMF–Government of Samoa Conference on Pacific Island Countries, held March 23, 2012, in Apia, Samoa. The authors thank Hoe Ee Khor, Tobias Haque, Matt Davies, Patrizia Tumbarello, Alexander Pitt, Jim Walsh, Nghi Luu, Craig Fookes, Ian Davidoff, Biman Chand Prasad, and Tiru Jayaraman for helpful comments. Tobias Haque kindly shared the initial World Bank database that is used in the econometric analysis, and Jan Gottschalk provided helpful input in the formative stage of the research.


The focus of this study is the 11 PICs that are IMF members: Fiji, Kiribati, Marshall Islands, Micronesia, Palau, Papua New Guinea, Samoa, the Solomon Islands, Tonga, Tuvalu, and Vanuatu. Timor-Leste is not included in this study.


See the forum communiqué of the Forty-Second Pacific Islands Forum, Auckland, September 7–8, 2011, and Noble, Pereira, and Saune (2011).


The definition of LICs is based on the current IMF classification. It should be noted that, as pointed out by Winters and Lim (2010), the growth performance of LICs tends to be understated, as some countries that were once LICs moved out of the group because of their faster economic growth. The classification of small states is based on Commonwealth Secretariat and World Bank (2000), with a population below 1.5 million except for Papua New Guinea. Of these countries, Annex 4.1 lists those with data available for econometric analysis. The Eastern Caribbean Currency Union members are Anguilla, Antigua and Barbuda, Dominica, Grenada, St. Kitts and Nevis, St. Lucia, and St. Vincent and the Grenadines.


Despite the food and fuel price shocks and the global financial crisis, average growth during 2007–10 was similar to that during 2001–06.


There is no consensus on the definition of small states—not only have several variables (for example, population, GDP, and geographic area) been used as a criterion, but also various thresholds for the same variable. The threshold for population, the most commonly used variable, has ranged widely from about 1 million to 15 million (see Armstrong and Read 2003). The IMF defines small developing states as those with a population of 1.5 million or less, excluding advanced economies (as defined by the IMF World Economic Outlook) and fuel-exporting countries.


The literature also discusses the potential advantages of small size. These include flexibility in adapting to changes in the external environment, a more homogeneous population, size-induced greater openness, and a lower chance of implementing costly import substitution. See Streeten (1993), Easterly and Kraay (2000), and Armstrong and Read (1998, 2003).


See McCallum (1995), Wei (1996), and Anderson and van Wincoop (2004) for the effect of international borders on trade. McCallum found that Canadian provinces trade up to 22 times more with each other than with states within the United States.


Milner and Westaway (1993) explore the medium-term growth effects of country size through possible capital shallowing, restricted structural change, barriers to catching up, and limited domestic technology diffusion. The authors find some evidence of capital shallowing and greater barriers in technological diffusion.


Redding and Venables (2004) find that the geography of access to markets and sources of supply is an important determinant of income levels across countries. Their results indicate that halving a country’s distance from all of its trade partners increases its income per capita by about 25 percent. Similarly, using data on costs of doing business, Winters and Martins (2004) show that for both clothing and electronics assembly, microeconomies have cost inflation factors of 36 percent, and that for tourism the factor is 58 percent. The last is driven substantially by high costs for personal travel (and the high share of such travel in overall packages).


Srinivasan (1986) argues that smallness is neither a necessary nor sufficient condition for slow growth and development. Milner and Westaway (1993) find there is no obvious link between medium-term growth performance and a range of attributes of country size.


Gibson and Nero (2007) also find that long-term growth in small states is also hurt by output volatility and language diversity.


Most studies in the literature assume a same value of g + δ for all countries, and the most frequently used value is 0.05. See, for example, Mankiw, Romer, and Weil (1992), Islam (1995), Yao and Zhang (2001), and Chen (2012).


In our panel data exercise, the whole sample period 1992–2008 is divided into four: 1992–96, 1996–2000, 2000–04, and 2004.


To test investment endogeneity, manufacturing and mining output relative to GDP and agricultural output relative to GDP were used as external instruments for the investment-to-GDP ratio. The Sargan test statistic is 0.811 with a p-value of 0.617, therefore the null hypothesis that external instruments are valid is not rejected. However, the Hausman test statistic of 1.618 with p-value of 0.203 provides strong evidence that the null hypothesis of exogeneity is not rejected.


For the aid-to-GDP endogeneity test, we considered external instruments such as the education index, deaths caused by natural disasters, government expenditure to GDP, and primary industry output to GDP. Only government expenditure to GDP and primary industry output to GDP are found to be valid external instruments based on the Wald parameter test (F-statistic = 13.45). Overidentification of parameters is confirmed by a Sargan chi-squared statistic of 0.453 with p-value of 0.5011. However the Durbin-Wu-Hausman test (with a chi-squared statistic of 0.46481 and p-value of 0.495) provides no evidence that the aid-to-GNI ratio in the analysis is endogenous. Nonetheless, the two-stage least squares estimator was applied to address any potential endogeneity, but the results obtained are similar to those using the ordinary least squares estimator.


To test for endogeneity of growth volatility, we used political stability and the exports-to-GDP ratio as external instruments. The Sargan test statistic of 0.111 with p-value of 0.739 strongly indicates that external instruments are valid, and the Hausman test statistic of 9.316 with p-value of 0.002 provides strong evidence that volatility is endogenous.


We also tried a specification with human capital as a growth determinant. We used various proxies for human capital, such as the United Nations’ education and human development indices and the secondary school enrollment rate, but none were found statistically important in explaining growth experience in small states. Data limitation in terms of time coverage may have affected the results.


Discussions hereafter are based on regressions 2 and 3 in Table 4.1.


Small states in Africa also suffer a geography-related disadvantage, but PICs are subject to by far the greatest disadvantage among small states. We run two models: one using trade-weighted distance and the other regional dummies to capture the impact of geography on growth. Not surprisingly, we found the two variables are correlated, and including one in the regression would render the other insignificant. Moreover, the estimated impact of the two sets of variables on PIC growth is similar, confirming that distance is the main obstacle to growth. Distance is measured as GDP-weighted physical distance to the capital cities of major markets.


We did include secondary education in the regression in an attempt to capture the effect of human capital on growth. However, the results are statistically insignificant.


This result is illustrative only. Our estimates are significant only at the 83 percent confidence level, and the sample size is small (28 observations). However, the result is consistent with that reported in Easterly and Kraay (2000).


The relationship between aid and growth is a hotly debated subject in the Pacific, as it is globally. See Bowman and Chand (2008), Rao, Sharma, and Singh (2008), Pavlov and Sugden (2006), and Hughes (2003).


A growth diagnosis approach could be employed in such an analysis. See Duncan and Nakagawa (2006) for a growth diagnosis for six PICs.


The six are Fiji, Papua New Guinea, Samoa, the Solomon Islands, Tonga, and Vanuatu.


From April 2014, the kina (the currency of Papua New Guinea) has reduced its volatility and followed a trend within a 2 percent band against the U.S. dollar. Accordingly, the de facto exchange rate arrangement was reclassified from floating to a crawl-like arrangement, effective April 11, 2014.


The textiles, clothing, and footwear industry accounted for 26 percent of Fiji’s total domestic exports in 1997, contributing 3.5 percent of GDP and providing employment for about 18,000 people (16 percent of those in total paid employment). In 2010, the industry reportedly employed only 4,000 people.


Negotiations on PACER-Plus were started under PACER in 2007.


Data are available only for Fiji, Papua New Guinea, Samoa, the Solomon Islands, Tonga, and Vanuatu.


IMF (2011) discusses the growing economic ties between low-income countries and the BRICS (Brazil, Russia, India, China, and South Africa).


Other PICs exclude Papua New Guinea and the Solomon Islands from the group of PICs used in this study.


Using a panel regression, Thacker and Acevedo (2011) find a significant positive association between tourism and growth. A 10 percent increase in tourist arrivals per capita raises economic growth by about 0.2 percent.


Note that the weight-based measures tend to overstate the decline in transport costs because the composition of shipped goods tends to shift toward lighter, higher-value goods over time.


Kleinert and Spies (2011) find that trade partners with 10 percent more exports enjoy 0.8 percent lower transportation costs. Favaro, Halewood, and Rossotto (2008) show how the lack of competition in the telecommunications industry in Samoa raised costs to consumers.


Chapter 12 discusses macroeconomic policy coordination in the context of strengthening monetary policy transmission.


One thinks of the Chinese experience in recent years—the yuan has faced pressure for appreciation despite China having higher inflation than its trading partners.


Chen and Singh (2011) estimate that total factor productivity growth in Fiji during 1983–2007 was only 0.5 percent a year, less than half the rates in Asia and other Pacific countries.


Pacific Forum Line was a multi-government-run shipping line. It was born of PICs’ concern over the deterioration of traditional island tramp services due to containerization. Samoa recently acquired all outstanding PIC shares in Pacific Forum Line.


The Pacific Islands Telecommunications Association is a nonprofit organization formed to represent the interests of small island nations in the Pacific region in the field of telecommunications.


United Nations–defined least-developed countries in the Pacific include Kiribati, Samoa, the Solomon Islands, Tuvalu, and Vanuatu.

    Other Resources Citing This Publication