Resilience and Growth in the Small States of the Pacific

Chapter 9. Vulnerabilities of Isolated Small Island States in the Pacific

Hoe Khor, Roger Kronenberg, and Patrizia Tumbarello
Published Date:
August 2016
  • ShareShare
Show Summary Details
Chris Becker

This chapter considers features of small island states in the Pacific region as well as in Asia, and documents some of the key characteristics that set them apart from small states in other regions. The chapter restricts itself to a limited number of general indicators that are largely macro oriented. In particular, it considers population size, income per capita, the fertility of the land, and the ability to tap into economies of scale. It captures the degree of geographic isolation confronted by some countries. As a result, we leave aside equally important but more micro-oriented variables, such as telecommunications or electricity generation, as well as development indicators, such as literacy or infant mortality rates.

We show that small island states in the Pacific are typically different from small states in other regional groupings in that they are extremely geographically isolated and have limited scope to tap into economies of scale due to their small populations. The degree of smallness can complicate the interpretation of income per capita as a proxy for economic conditions, and the lack of fertile land constrains opportunities for some countries to diversify food sources away from imports. The cursory empirical evidence presented appears consistent with earlier evidence that most of these factors are important determinants of economic outcomes in small states.

Small states in other parts of the world have their own characteristics that lead to a somewhat different set of factors that might be associated with vulnerability to shocks. These may or may not have been taken into consideration in this chapter, which employs a far less than exhaustive set of potential indicators. So while the scope is purposely limited to the countries of principal interest for this study, it could provide a basis for more comprehensive future research.

The rest of the chapter discusses a number of key macro-related variables in the context of small states. These cover the definition of small states in terms of population size, constraints on the ability to take advantage of economies of scale, income per capita as a measure of poverty, the relationship between land fertility and import replacement through agriculture, and, perhaps most important, the degree of geographic isolation. These indicators are summarized by a simple and transparent ranking, and an attempt is made to relate the rankings to actual economic outcomes.

Defining Small States by Population

There is no universally accepted definition of what makes a country small. Relevant metrics include population, size of the land or territory (including maritime zones), and gross national income (GNI).1 However, since population is usually correlated with other variables such as GNI in most countries, the number of residents is often used as the measure that defines smallness. The World Bank defines a small state as one with a population of less than 1½ million people (World Bank 2007), but it also notes that no definition, whether it be population, geographic size, or income, is likely to be fully satisfactory. In practice, any threshold used has an arbitrary element, and larger states that lie outside this definition will share some of the characteristics or vulnerabilities of smaller countries.

Using World Bank data, we are able to derive a consistent data set for a preliminary analysis of a sample of 50 small states (see Annex 9.1).2 The smallest of these is Tuvalu in the Asia and Pacific region, with a population of only about 10,000 people. The largest is Trinidad and Tobago in the Caribbean, with some 1.3 million people. Figure 9.1 ranks these countries by population. Those in the Asia and Pacific region are highlighted in a darker color.

Figure 9.1Small States with Populations of Less than 1.5 Million, 2013

(Population, thousands)

Source: World Bank, World Development Indicators database.

1 The figure for Gibraltar is from 2012.

Of the 10 smallest states, four are in Europe and central Asia, three are small island states in the Pacific (Marshall Islands, Palau, Tuvalu), and two are in Latin America and the Caribbean. Greenland represents an outlier not easily classified in any region.3 The difficulty of defining aspects of smallness is usefully illustrated by examining these extremely small countries (sometimes referred to as microstates). While the number of people residing in them is indisputably small, they can be quite heterogeneous in terms of other economic indicators that make differentiation between small states important. For example, while the population of Monaco is about three times larger than that of Tuvalu, its GNI per capita is almost 40 times larger. Extremely small populations are therefore not always systematically related to income, but in some cases raise a number of other relevant idiosyncratic considerations. As a result, of the 10 smallest countries in this sample, fewer than half are eligible under the income criteria for both concessional borrowing from the International Development Association (IDA) (2001) and the IMF’s Poverty Reduction and Growth Trust.4

Lacking Economies of Scale

Generally speaking, the relationship between population and total income appears to be positive. This is not surprising given that labor is a major input into the production process and a larger workforce can produce a greater number of goods and services for higher total income. Figure 9.2 shows this relationship for the small state sample and plots a simple trend line, which, admittedly, has only weak explanatory power. Asia and the Pacific is again highlighted by a darker color.

Figure 9.2Relationship between Population and GNI

(GNI, millions of U.S. dollars)

Sources: World Bank, World Development Indicators database; and author’s calculations.

Note: GNI = gross national income.

Since population is only one of many factors that determine income, the distribution of countries around the trend is relatively wide. Indeed, there appears to be a tendency to deviate further as population increases. Several outliers to the relationship are notable, with Luxembourg generating the largest GNI despite a relatively small population.5 There is a confluence of attributes that makes some countries able to overcome the barriers of having a small resident workforce. In the case of Luxembourg, specialization in the provision of high-value-added financial services is possible due to a number of unique features, including very close proximity to highly skilled labor markets and infrastructure in neighboring countries, notably France and Germany. A significant share of the people employed in Luxembourg do not reside in the country itself, but commute daily from neighboring towns across the national border. Similarly, centers for trade and commerce such as Macao

Special Administrative Region are able to generate higher income than countries of similar size with less of this type of economic activity. The key feature of such trade hubs tends to be strategic location on major trade and shipping routes that act as gateways to major markets.

The small island states in the Pacific tend to lack the attributes enjoyed by financial centers and trade hubs. They are typically considerable distances from the major labor markets that are better equipped to supply a more skilled workforce, such as Australia and New Zealand (and this is discussed further later in the chapter). Geography also precludes sharing infrastructure with more developed neighbors. Indeed, basic infrastructure may need to be duplicated if populations are dispersed, as they are in countries comprising groups of islands. Furthermore, they are usually not on the routes of shipping lanes, which connect major producers with markets and are therefore unlikely candidates for establishing trade hubs. Nonetheless, there are certain types of financial services that are less closely linked to geographic location, such as business registration and the incorporation and registration of international cargo ships, which small Pacific islands are not precluded from pursuing.

Fixed Costs Spread over a Narrow Base

One implication of small population is that total income, and therefore the tax base, might be so small that the fixed costs associated with providing public goods and services have to be spread across a very narrow base. As a result, fixed costs may represent an unusually high share of national income. One consequence might be that there is insufficient tax revenue to secure the public provision of basic health services, transportation, and government administration. These constraints often result in small administrations that lack the capacity to function efficiently. The degree to which this is an issue varies among countries.

This is an argument related to economies of scale in that the provision of public goods and services might not be possible at the quantity necessary to minimize the average cost, giving rise to inefficient outcomes that are most difficult to successfully deal with in extremely small countries. Another consideration might be the expense incurred in building infrastructure resilient to frequent natural disasters and adverse weather conditions.

One simple way to make an assessment about the available opportunities to reap economies of scale is to calculate the percentage deviation of each country’s GNI from the median of the small state sample.6 This allows calibration of which states are the smallest and therefore the most likely to be lacking in the ability to exhaust economies of scale. The calculations are shown in Table 9.1. While very small states such as Palau and Tuvalu are more than 90 percent smaller than the sample median, others are considerably better placed. It is notable that of the small countries in the Pacific, almost all are so small that the scope to access economies of scale could be a significant issue. Only Fiji, Macao Special Administrative Region, and Timor-Leste were larger than the median income of $1.7 billion in 2010.

Table 9.1Percentage Deviation from Small State Median Income
StatePercent BelowPercent Above
San Marino−8.5
Marshall Islands−88.6
St. Kitts and Nevis−64.2
Isle of Man131.1
Antigua and Barbuda−32.0
St. Vincent and the Grenadines−60.0
Channel Islands495.8
St. Lucia−33.6
São Tomé and Príncipe−88.4
The Bahamas305.7
Brunei Darussalam624.9
Cabo Verde−5.8
Solomon Islands−67.9
Macao SAR977.9
Equatorial Guinea492.3
Trinidad and Tobago1,099.9
Sources: World Bank, World Development Indicators database; and author’s calculations.Note: Median gross national income of small state sample was US$1,719 million, 2010. States are ordered by population.
Sources: World Bank, World Development Indicators database; and author’s calculations.Note: Median gross national income of small state sample was US$1,719 million, 2010. States are ordered by population.

Scope for Regional Cooperation

Notwithstanding the country-specific constraints on accessing economies of scale, there may be scope to improve the position of small states through regional cooperation (see also Hausmann 2001). A number of small states could come together to form a larger common market for goods and services, or share access to certain types of infrastructure. Figure 9.3 sums the population of all countries in each region to derive a regional aggregate as an indication of the potential for cooperation.

Figure 9.3Small States: Population by Region

Source: World Bank, World Development Indicators database.

Note: Population measured as the sum of population in all states in the regions shown.

In terms of the scope for cooperation in building scale based on the aggregate number of people in each region, the Pacific islands do not appear to be at a disadvantage relative to other regions. Instead, it would seem that small states in south Asia, as well as in the Middle East and North Africa, are the most limited in tapping into the benefits from cooperation—assuming that they are also somehow unable to cooperate with larger neighbors.7 Greenland may be of less concern owing to its special relationship with Denmark.

Nonetheless, there may be many other obstacles to effective cooperation. Some of these might be related to culture, language, distance between countries, legal structure, and political and other forms of heterogeneity among neighboring states, which will differ according to region. However, there are regional forums that facilitate policy coordination and discussion, and this has certainly been the case for some time in the Pacific.8

Gauging Vulnerability from Income Per Capita

Income per capita is an important variable, not least because it is often used as a proxy indicator for poverty and aid eligibility. For example, IDA eligibility is principally determined by a threshold related to GNI per capita, which is reviewed periodically. Eligibility for the IMF’s Poverty Reduction and Growth Trust takes its cue from the IDA and is therefore similarly based on income per capita.

However, income per capita might also suffer from being an imperfect indicator, in the same way that the separate consideration of total income and population in relation to the possible inability of very small states to access economies of scale are imperfect indicators. Take, for example, an extremely small state. It might generate sufficient income from the sale of fishing licenses and remittances to rank relatively high on income per capita because the total income numerator is shared across a very small population denominator. Even so, this state might find itself in a situation in which total income is so small that it proves prohibitively expensive for the government to provide adequate health services domestically. If a substantial part of that higher income per capita has to be allocated toward expensive medical services in another country, the remaining disposable per capita amount available to households for consumption and saving might be significantly lower. In this simple example it is easy to recognize that assessments based on income per capita alone can be misleading.9 This problem becomes more pronounced as the denominator in the calculation takes an extremely small value.

Figure 9.4 plots the income per capita for the sample of 50 small states. One consideration is that while this sample is restricted to small states, there are many countries with populations that exceed the 1½ million threshold but have income per capita that is much smaller than several of the countries depicted. This is typically attributable to a very small numerator relative to a very large denominator in the calculation. It therefore becomes important to understand the determinants of income per capita when making policy decisions.

Figure 9.4Small States: Gross National Income per Capita, 2010

(Current U.S. dollars)

Source: World Bank, World Development Indicators database.

The top decile of five small states with the highest GNI per capita is entirely composed of European countries (Channel Islands, Liechtenstein, Luxembourg, Monaco, San Marino). The proximity to neighboring countries—and access to their workforce, infrastructure, and the nearest continent—may well be important explanatory factors in the scope to overcome disadvantages associated with being small (Martins and Winters 2004). The proximity to neighboring countries also raises the need for consideration of multiple vulnerabilities for some small states. While it might be feasible to routinely overcome one or even two types of exposure, once countries are disadvantaged by several such factors, it becomes less likely that they can overcome all of them all the time and escape the adverse consequences of their vulnerabilities.

Scope for Import Substitution Through Agriculture

Another common feature of small states is that, despite sometimes very large overall land masses, they typically have very little fertile or arable land available for cultivation.10 One reason why this might be considered an important macro-related indicator for countries is because it could be related to the ability to substitute for imports of foodstuffs. Countries may well be more exposed to balance of payments crises if they depend heavily on food imports because of inadequate conditions to foster domestic agriculture. The most striking example of this is Greenland, where an extremely large landmass is very sparsely populated and there is no arable land (Table 9.2). The vulnerabilities resulting from excessive dependence on imports of foodstuffs were sharply underscored in recent years when food prices rose sharply (see Chapters 6 and 7).

Table 9.2Scope for Agriculture
StateTotal Landmass (km2)Nonarable Land (km2)Arable Land (km2)Arable Land (m2 per Capita)
San Marino605010318
Marshall Islands18016020380
St. Kitts and Nevis26022040738
Isle of Man57052050582
Antigua and Barbuda44036080889
St. Vincent and the Grenadines39034050457
Channel Islands19015040247
St. Lucia61058030165
São Tomé and Príncipe960860100518
The Bahamas10,0109,93080212
Brunei Darussalam5,2705,2403072
Cabo Verde4,0303,4306001,203
Solomon Islands27,99027,830160285
Macao SAR282800
Equatorial Guinea28,05026,73013201,744
Trinidad and Tobago5,1304,880250186
Source: World Bank.Note: Countries are ordered by population, as of 2013. Gibraltar’s population data are for 2012; km2 = square kilometer; m2 = square meter.
Source: World Bank.Note: Countries are ordered by population, as of 2013. Gibraltar’s population data are for 2012; km2 = square kilometer; m2 = square meter.

In terms of the square meters per capita available for cultivation, some countries are in a considerably worse position than others. Five states in the sample countries have no arable land at all (Gibraltar, Greenland, Macao Special Administrative Region, Monaco, Tuvalu). Several Pacific islands suffer from very infertile soil or, in some cases, no soil at all. Many consist of nothing more than coral or sand, and Tuvalu stands out as being particularly infertile and unsuitable for agriculture. On the other hand, some of the larger Pacific islands, such as Fiji, Samoa, Timor-Leste, and Tonga, are relatively fertile, especially when compared to some states in Africa. On average, the states in the Pacific may not be as infertile as desert countries in the Middle East and North Africa or Greenland, but they are notably more infertile than small countries in Europe, Latin America, and the Caribbean (Figure 9.5).

Figure 9.5Small States: Fertility of Land by Region

(Average square meters per capita)

Sources: World Bank, World Development Indicators database; and author’s calculations.

The regional findings in Figure 9.5 are not surprising given the geographic location of states in the sample. European countries stand out as having the most arable land and, with the significant exceptions of India and Nigeria, the rest of the world has less agriculturally productive land (Figure 9.6).11 There are, of course, several smaller exceptions to this.

Figure 9.6Arable Land

(Percent of total landmass)

Source: U.S. Central Intelligence Agency, The WorldFactbook.

Trade with highly agriculturally productive countries is one way to overcome the domestic constraints that some small states face. Some of the important factors in determining that possibility are access to alternative resource endowments that can be traded and the proximity to trading partners. While Pacific island states might be able to trade fish stocks, or use the revenue from selling fishing rights, they remain disadvantaged by their geographic isolation, which raises transport costs and can even prove prohibitive to gaining market access.

Geographic Isolation

The most distinguishing characteristic of small Pacific islands is how remote they are, not only from the nearest continent, but also from neighboring countries. While technological progress has allowed countries to overcome barriers such as those related to effective communication, distance remains a key challenge to overcome when physical factors are important. The transportation costs associated with trade and commerce are therefore commensurately higher as distance increases (Commonwealth Secretariat and World Bank 2000; Zhu 2012).12 This problem is compounded by import dependence, especially for foodstuffs, due to nonarable land, as discussed earlier. Martins and Winters (2004) show that small economies might not even be suitable locations for tourism unless they have specific comparative advantages that allow them to charge substantially higher prices to overcome cost disadvantages. Furthermore, since this geographic isolation is closely associated with the dispersion of many small islands in the Pacific Ocean, there is also a link to the susceptibility of these states to natural disasters such as tsunamis and hurricanes. The environmental challenges also extend to issues associated with rising sea levels and global warming, although some of these are in common with other regions. This is especially so in the Caribbean, where small states are vulnerable to similar environmental pressures. Annex 9.1 lists some relevant indicators on this issue, such as the distance to closest continent subindex of the Environmental Vulnerability Index calculated by the South Pacific Applied Geoscience Commission and the United Nations Environment Programme.

In the sample of small states, Kiribati, Marshall Islands, Samoa, Tonga, and Tuvalu are some of the most isolated states in the world (Figure 9.7). Each is more than 3,000 kilometers (1,900 miles) from the nearest continent, Australia.

Figure 9.7Small States: Distance to Nearest Continent

(Kilometers; countries ordered by population)

Source: South Pacific Applied Geoscience Commission.

The vulnerability represented by their geographic isolation therefore notably differentiates small islands in the Pacific from many of the other small states in the sample. The average distance to the nearest continent for Pacific islands is more than four to five times that applicable to the average country in the Caribbean or sub-Saharan Africa (Figure 9.8). On the other hand, small states in Europe, northern Africa, and the Middle East are considerably less isolated on the measure used here.13

Figure 9.8Small States: Distance to Nearest Continent by Region

(Average kilometers)

Source: South Pacific Appled Geoscience Commission database.

When considering the issue of isolation, additional factors beyond the scope of this chapter are nevertheless worth mentioning briefly. Mayer and Zignago (2011) calculate a more comprehensive measure of remoteness by including not just the single distance between a country and the nearest continent, but by measuring the distance between each country and all other countries in their sample of 224 countries. This metric lends further support to the finding that small Pacific island states are particularly isolated. The main driving factor is that these states are not only far away from the nearest continent, but are widely dispersed over a vast area of ocean and therefore also very far away from each other and all other countries.14

An interesting further augmentation of the data is to weight these distances by GDP to capture how physically far removed countries are from major world markets (see Chapter 4). Once again, this augmentation makes small states in the Pacific even worse off. Even though Australia has a very large landmass (the sixth largest state in the world), it has a relatively small population and therefore also represents a much smaller market than large neighbors, such as China, the euro area, or the United States, to some small states. This consideration of distance from major markets is also relevant to some countries in Africa even though the state might be on the actual continent, and could change some of the results just discussed.

Vulnerability Ranking

A simple way of summarizing this type of information is to rank states according to how they are positioned relative to other small states on the factors discussed in this chapter. The aim is to keep the summary indicator as simple and transparent as possible. We note the trade-offs involved from the outset. Mechanical indices can never fully reflect the complexities and changing dynamics involved in the interaction between these variables. Additionally, there are limitations that arise from the inputs into the calculations being far from exhaustive in their description of small states. Nevertheless, we hope to convey some of the key characteristics that set small Pacific islands apart from small states in other regions.

Ranking Small States

We use each variable discussed in the chapter and calculate how every small state ranks relative to all others. For example, if we rank the sample of 50 states according to population size, one might consider the smallest state as being the most vulnerable. The country with the smallest population, Tuvalu, is given an index ranking of 50 and the country with the largest population, Trinidad and Tobago, is given an index ranking of one. Larger numbers therefore indicate greater relative vulnerability on the indicator in question.15 Similarly, countries with the least amount of arable land per capita might be vulnerable, as would be the countries that are the farthest away from the nearest continent and are the most isolated. Countries with the smallest absolute U.S. dollar level of GNI are probably less able to reap economies of scale in the provision of public goods and services and could be disadvantaged. Similarly, those with the lowest income per capita might typically be considered to be relatively poor and therefore exposed to adverse shocks that cannot be easily absorbed without assistance from the international community.

We try to capture vulnerability by synthesizing the measures just discussed into an index. An aggregate summary ranking is achieved by calculating the equally weighted average across the five individual indicator rankings used in this simple study.16 The result is a broad reflection of which states in the sample are the most vulnerable. Table 9.3 provides the details of the calculations and ranking.

Table 9.3Small States Ranked by Indicator
PopulationArable LandDistanceGNIGNI per CapitaAverage
Marshall Islands443749483542.6
St. Kitts and Nevis433042392235.2
São Tomé and Príncipe282224474833.8
Solomon Islands162041404933.2
St. Lucia303334322831.4
Antigua and Barbuda372543312031.2
St. Vincent and the Grenadines322827373030.8
San Marino47421127526.4
Cabo Verde191431263825.6
Brunei Darussalam21353671122.0
Isle of Man39291917622.0
Channel Islands3132209419.2
The Bahamas222423131419.2
Macao SAR1548166818.6
Trinidad and Tobago1161831711.0
Equatorial Guinea1175101810.2
Source: Author’s calculations.Note: Vulnerability is ranked on a scale of 1 to 50. GNI = gross national income.
Source: Author’s calculations.Note: Vulnerability is ranked on a scale of 1 to 50. GNI = gross national income.

According to this metric, Tuvalu is the most vulnerable small state in the sample. It has a very small population, no arable land, is very isolated, and has little scope for accessing economies of scale in the provision of public goods and services, on account of its small GNI. These factors more than offset its more favorable ranking on income per capita.

Estonia is at the other extreme. In the context of small states, it has a relatively large population, its land is very fertile, it is surrounded by close European neighbors, and total GNI is quite high. These factors more than offset a fairly low ratio of income per capita. This outcome is not intended to imply that Estonia, or other states in the sample, do not face other substantial vulnerabilities. Instead, it is simply a reflection of how states compare based on just the five indicators chosen to illustrate the relative position of small states in the Pacific. Broadly speaking, some tentative conclusions can be reached about the average regional characteristics of small states (Figure 9.9).

Figure 9.9Regional Vulnerability by Indicator

(Index rank)

Source: Author’s calculations.

Note: GNI = gross national income.

Small states in the Pacific are the most vulnerable on a number of counts considered in this chapter. This is in part driven by several common vulnerabilities such as isolation, but also by the extreme exposure of some states in the region to several additional indicators of vulnerability. These states also lack the ability to reap economies of scale, generally have low income per capita, small populations, and little arable land. Compared with other regions, they rank worse than the average of 25.5 on all measures considered.17 Small states in Latin America and the Caribbean, south Asia, and sub-Saharan Africa are probably somewhat less vulnerable to the factors considered here. These regions appear to have a common degree of overall exposure. Small countries in Europe appear to be the least disadvantaged in this sample—a result in large part driven by a number of outliers that are highly developed and rich countries that happen to have small populations but do not appear particularly disadvantaged by this characteristic. It is therefore worth keeping in mind that other small European countries are less fortunate, and that for the Caribbean, south Asia, and Africa we might need to improve our means of identifying and capturing other forms of vulnerability.

Empirical Link to Real Economic Outcomes

A series of simple linear regressions are fitted to conduct a preliminary investigation of the relationship between the potential vulnerability indicators for small states and real economic outcomes. The growth rate in GNI is therefore the dependent variable we are trying to explain using indicators of vulnerability.

Economic outcomes are proxied by annualized nominal growth in GNI measured in U.S. dollar terms between 2001 and 2010. The data are annual and therefore only nine observations are available for most states. Three states (Montenegro, São Tomé and Príncipe, Timor-Leste) have fewer observations than this, and Gibraltar had to be dropped from the sample of 50 states because of a lack of time series data. We also note the high likelihood of cross-correlation between growth outcomes during the financial crisis, given that it represents a common shock to all countries in the sample, albeit with different intensities.

To investigate the usefulness of ranking countries by their degree of vulnerability, we fit separate regressions using each of the five vulnerability indicators shown in Table 9.3: population, arable land, distance, income, and income per capita (see also Gallup, Sachs, and Mellinger 1998; Kuznets 1960). Since states are ranked according to their relative degree of vulnerability in each of these indicators, we attempt to capture the relationship between growth and vulnerability. Our prior is to find a negative relationship between the relative degree of vulnerability exhibited by a state and the average growth rate it is able to achieve. For example, we would expect that a high ranking on distance—which by construction indicates that a state is relatively isolated—results in lower growth outcomes than for states that are less isolated.

A closer look at the regressions (Table 9.4) indicates that the slope coefficient for almost all individual indicators is consistent with the expected negative relationship between relative vulnerability and growth. A notable exception is income per capita, which indicates a weak positive relationship with growth. Possible explanations for why lower income per capita might be associated with faster growth could relate to structural factors such as developing economies typically being able to grow more rapidly than more developed economies (which would also tend to have higher income per capita). More realistically, the relationship between income per capita and growth is probably not a very meaningful indicator, especially in the case of microstates, and this is reflected in a statistically insignificant relationship as indicated by the p-value on the variable.18

Table 9.4Relationship between Vulnerability Ranking and Growth
Ranking According to the ith Indicator of Vulnerability Number1Slope Coefficientp-valueRMSE(ith)RMSE(ith)/RMSE(index)
Arable land−0.160.0063**5.401.0651
Income per capita0.020.68995.841.1519
Memorandum item:
Source: Author’s calculations.Note: RMSE = root mean square error. * Denotes significance at the 5 percent level; ** significance at the 1 percent level.1 Sample excludes Gibraltar owing to data availability.2 Additional exclusions: (i) income per capita in the calculation of the Index, (ii) outliers (Timor-Leste and Equatorial Guinea).
Source: Author’s calculations.Note: RMSE = root mean square error. * Denotes significance at the 5 percent level; ** significance at the 1 percent level.1 Sample excludes Gibraltar owing to data availability.2 Additional exclusions: (i) income per capita in the calculation of the Index, (ii) outliers (Timor-Leste and Equatorial Guinea).

Among a number of other interesting findings is that the most statistically significant (p-value) explanatory variables for growth are size of the population, the degree of isolation measured by distance to the nearest continent, and fertility of the land.19 As a result, the ratio of the root mean square error relative to the benchmark model fitted for the overall average index is lowest for these three core indicators. Furthermore, the combination of all indicators into the summary index yields the best fit. In part, this is because more variation and information are reflected by the index to explain the dependent variable, but we would also argue that the combination of vulnerabilities is important in influencing economic outcomes. It is clearly more difficult to register consistently good economic performance when exposed to a significantly larger number of sources for adverse shocks. This relationship is plotted in Figure 9.10. As in previous figures, states in Asia and the Pacific are depicted in a darker color.

Figure 9.10(Index rank)

Vulnerability and Growth

Source: Author’s calculations.

Note: GNI = gross national income.

The fitted relationship indicates the expected negative relationship between the degree of vulnerability and growth outcomes for small states, but it is significantly affected by two outliers on growth (Equatorial Guinea, Timor-Leste).20 One interpretation of the clustering of Pacific states in the top-left quadrant of Figure 9.10 would be that their relatively high degree of vulnerability does indeed impede their economic performance by dragging average growth lower.

As a final illustration, we refit the equation using a recalculated average ranking index that excludes income per capita on the basis that its explanatory value was found to be statistically insignificant, and drop the two outliers from the sample.21 The results are shown in Table 9.4 as index-ex. Not surprisingly, the fit improves dramatically. The root mean square error is significantly smaller and improves on the aggregate average index of vulnerability by more than 40 percent in explaining average growth outcomes.22


This chapter does not definitively find that small states in the Pacific are absolutely more vulnerable than small states in other regions—the scope is simply too narrow to address this question adequately. It does, however, show that on the limited number of macro-oriented indicators considered, most small Pacific island countries rank as being particularly exposed to adverse shocks relative to their peers in other regions. We find that population size, distance from the nearest continent, arable land, and scope to exploit economies of scale are all statistically significant in explaining economic outcomes in small states. The combination of these vulnerabilities into an overall index lends support to the idea that a confluence of vulnerabilities is also important in determining growth outcomes.

Small island states in the Pacific are disadvantaged because they are sometimes extremely small in terms of population and, consequently, limited in being able to access economies of scale in the production of goods and services. The quality of soil is often not very good, and, as a result, they can be exposed to the disadvantages that follow from being heavily dependent on imports of foodstuffs.

It is, however, the extreme degree of isolation of most small states in the Pacific that is quite unique in defining their vulnerabilities on several facets. While technological progress has helped to bridge the communication chasm, when it comes to physical considerations in areas such as trade, commerce, and labor mobility, significant barriers with economic consequences are not only important, but also unlikely to be resolved in the foreseeable future. Furthermore, as a result of their geography, small states in the Pacific face a number of environmental challenges such as tsunamis, hurricanes, and rising sea levels. The tyranny of the sheer distances is therefore likely to remain a key challenge in the Pacific and will remain at the forefront on the minds of informed policymakers for some time.

Annex 9.1. State Classification Data
Annex Table 9.1.1Summary Statistics
Country NameRegional ClassificationPopulation as of 2013 (Thousands)Arable Land as of 2013 (m2 per Capita)EVI-13 Isolation Index (1–7)Distance to Closest Continent (km)Closest ContinentGNI Current Prices as of 2010 (U.S.$ millions)GNI/Capita Current Prices as of 2010 (U.S.$)GNI/Capita PPP Basis as of 2010 (U.S.$)
TuvaluEast Asia and Pacific10073,300Australia474,7605,351
PalauEast Asia and Pacific2147861,463Australia1346,56011,000
GibraltarEurope and Central Asia29010Europe96933,15133,151
San MarinoEurope and Central Asia3131810Europe1,57249,86635,163
LiechtensteinEurope and Central Asia3781210Europe4,903137,070151,189
MonacoEurope and Central Asia3810Europe6,479183,150182,990
Marshall IslandsEast Asia and Pacific5338073,500Australia1973,6403,003
St. Kitts and NevisLatin America and Caribbean5473872,100South America61511,74015,850
GreenlandGreenland565750North America1,46626,02025,604
DominicaLatin America and Caribbean728335698South America4586,76011,990
AndorraEurope and Central Asia7912610Europe3,44741,75036,113
Isle of ManEurope and Central Asia86582230Europe3,97248,91047,926
SeychellesSub-Saharan Africa891125475Africa8459,76021,210
Antigua and BarbudaLatin America and Caribbean9088972,110South America1,16913,17020,240
KiribatiEast Asia and Pacific10219573,250Australia2002,0103,530
MicronesiaEast Asia and Pacific10419372,500Australia3042,7303,490
TongaEast Asia and Pacific1051,51973,200Australia3423,2804,580
GrenadaLatin America and Caribbean1061894140South America7246,9309,890
St. Vincent and the GrenadinesLatin America and Caribbean1094574300South America6886,30010,830
Channel IslandsEurope and Central Asia162247230Europe10,24167,96066,781
St. LuciaLatin America and Caribbean1821655698South America1,1426,56010,520
SamoaEast Asia and Pacific1901,31373,800Australia5493,0004,270
São Tomé and PríncipeSub-Saharan Africa1935184225Africa1991,2001,920
VanuatuEast Asia and Pacific25379171,800Australia6332,6404,320
BarbadosLatin America and Caribbean2855624350South America3,45412,66019,000
IcelandEurope and Central Asia32321761,000Europe10,38132,71027,680
BelizeLatin America and Caribbean3322,10910North America1,3133,8106,210
MaldivesSouth Asia3451165500Asia1,8185,7508,110
The BahamasLatin America and Caribbean3772124200North America6,97320,61024,800
Brunei DarussalamEast Asia and Pacific418726950Asia12,46131,80050,180
MaltaEurope and Central Asia4231894250Europe7,95819,27024,840
Cabo VerdeSub-Saharan Africa4991,2035580Africa1,6203,2703,820
SurinameLatin America and Caribbean5391,07610South America3,0775,9207,680
LuxembourgEurope and Central Asia5431,14110Europe39,03077,16061,790
Solomon IslandsEast Asia and Pacific56128571,950Australia5521,0302,210
Macao SAREast Asia and Pacific56622Asia18,52734,88045,220
MontenegroEurope and Central Asia6212,78410Europe4,2606,75012,930
ComorosSub-Saharan Africa7351,0894300Africa5507501,090
BhutanSouth Asia75499510Asia1,3611,8704,990
Equatorial GuineaSub-Saharan Africa7571,74410Africa10,18214,54023,750
GuyanaLatin America and Caribbean8005,25310South America2,1642,8703,450
DjiboutiMiddle East and North Africa8732310Africa1,1051,2702,460
FijiEast Asia and Pacific8811,81672,600Australia3,1233,6304,510
CyprusEurope and Central Asia1,141762360Asia23,65529,43030,300
Timor-LesteEast Asia and Pacific1,1781,4004600Australia2,4932,2203,600
SwazilandSub-Saharan Africa1,2501,40110Africa3,1192,6304,840
MauritiusSub-Saharan Africa1,29667161,400Africa9,9257,75013,960
EstoniaEurope and Central Asia1,3254,49910Europe19,37114,46019,760
BahrainMiddle East and North Africa1,3328215Asia19,71418,73024,710
Trinidad and TobagoLatin America and Caribbean1,341186220South America20,62515,38024,040
Sources: The South Pacific Applied Geoscience Commission; and World Bank.Note: GNI = gross national income; km = kilometer; m2 = square meter; PPP = purchasing power parity.
Sources: The South Pacific Applied Geoscience Commission; and World Bank.Note: GNI = gross national income; km = kilometer; m2 = square meter; PPP = purchasing power parity.
Annex 9.2. Data Sources and Metadata
  • Arable land is sourced from data provided by the World Bank, as of 2010.
    • ◦ Arable land includes land defined by the United Nations Food and Agriculture Organization as land under temporary crops (double-cropped areas are counted once), temporary meadows for mowing or for pasture, land under market or kitchen gardens, and land temporarily fallow. Land abandoned as a result of shifting cultivation is excluded.
  • Environmental Vulnerability Index, EVI-13, Distance to Closest Continent, is sourced from the South Pacific Applied Geoscience Commission and the United Nations Environment Programme (2005).
    • ◦ This indicator captures the proximity of a state to the nearest continent. Note that if a state is within a continent, this value is zero. Isolated states may have a greater risk of loss of ecosystem types and species during periods of stress if they are far away from refuges and sources of recolonization. Isolated states are also likely to support fewer species than those close to large continents or biogeographic centers of radiation. In addition, there is less chance of genetic interchange (part of genetic resilience) in isolated areas. The likelihood of isolation being an important part of a state’s ecological resilience would be especially important if there are interactions with ongoing human impacts. States close to sources of recolonization are likely to be less at risk of permanent species losses, compared with those far away, particularly if they are small or fragmented.
  • Gross national income (GNI), current U.S. dollars, is sourced from data provided by the World Bank, as of 2010.
    • ◦ GNI (formerly gross national product) is the sum of value added by all resident producers plus any product taxes (minus subsidies) not included in the valuation of output plus net receipts of primary income (compensation of employees and property income) from abroad. Data are in current U.S. dollars. GNI, calculated in national currency, is usually converted to U.S. dollars at official exchange rates for comparisons across economies, although an alternative rate is used when the official exchange rate is judged to diverge by an exceptionally large margin from the rate actually applied in international transactions. To smooth fluctuations in prices and exchange rates, a special Atlas method of conversion is used by the World Bank. This applies a conversion factor that averages the exchange rate for a given year and the two preceding years, adjusted for differences in rates of inflation between the state and, up to 2000, Group of Five states (France, Germany, Japan, United Kingdom, United States). Starting in 2001, these states include those in the euro area, Japan, the United Kingdom, and the United States.
  • GNI per capita, current U.S. dollars, is sourced from data provided by the World Bank, as of 2010.
    • ◦ GNI per capita (formerly gross national product per capita) is the gross national income, converted to U.S. dollars using the World Bank Atlas method, divided by the midyear population.
  • GNI per capita based on purchasing power parity is sourced from data provided by the World Bank, as of 2010.
    • ◦ This is GNI converted to international dollars using purchasing-power-parity rates. An international dollar has the same purchasing power over GNI as a U.S. dollar has in the United States.
  • Some values estimated by the author using GDP and GDP per capita data are from the U.S. Central Intelligence Agency, World Factbook.
  • Population is sourced from data provided by the World Bank, as of 2010.
    • ◦ Total population is based on the actual definition of population, which counts all residents regardless of legal status or citizenship—except for refugees not permanently settled in the state of asylum, who are generally considered part of the population of their state of origin. The values shown are midyear estimates.
  • Regional classification of states is sourced from data provided by the World Bank, as of April 2012.
    • ◦ Geographic classifications and data reported for geographic regions are only for low- and middle-income countries, sometimes referred to as developing economies. The use of the term is convenient and is not intended to imply that all economies in the group are experiencing similar development or that other economies have reached a preferred or advanced stage of development. Classification by income does not necessarily reflect development status.
  • Land area is sourced from data provided by the World Bank, as of 2010.
    • ◦ Land area is a state’s total area, excluding area under inland water bodies, national claims to continental shelf, and exclusive economic zones. In most cases the definition of inland water bodies includes major rivers and lakes.

    BeckerChris. 2012. “Small Island States in the Pacific: The Tyranny of Distance?IMF Working Paper 12/223 International Monetary Fund Washington.

    • Search Google Scholar
    • Export Citation

    BrowneChristopher. 2006. Pacific Island Economies. Washington: International Monetary Fund.

    Commonwealth Secretariat and World Bank Joint Task Force on Small States. 2000. “Small States: Meeting Challenges in the Global Economy.

    • Search Google Scholar
    • Export Citation

    EasterlyWilliam and AartKraay. 2000. “Small States, Small Problems? Income, Growth, and Volatility in Small States.World Development (28) 11: 201327.

    • Search Google Scholar
    • Export Citation

    GallupJohn LukeJeffreyD. Sachs and Andrew D.Mellinger. 1998. “Geography and Economic Development.” Working Paper 6849 National Bureau of Economic Research Cambridge Massachusetts.

    • Search Google Scholar
    • Export Citation

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

    • Search Google Scholar
    • Export Citation

    International Development Association. 2001. “IDA Eligibility, Terms, and Graduation Policies.”

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2009. “IMF Reforms Financial Facilities for Low-Income States.” Public Information Notice 09/94 WashingtonJuly29.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2012. “Eligibility to Use the Fund’s Facilities for Concessional Financing.” Policy Paper Washington.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2013a. “Macroeconomic Issues in Small States and Implications for Fund Engagement.” Policy Paper Washington.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund (IMF). 2013b. “IMF Executive Board Reviews Facilities for Low-Income States and Eligibility for Using Concessional Financing.” Public Information Notice 13/45 Washington.

    • Search Google Scholar
    • Export Citation

    KuznetsSimon. 1960. “Economic Growth of Small Nations.” In Economic Consequences of the Size of Nations edited by E. A. G.Robinson. London: MacMillan.

    • Search Google Scholar
    • Export Citation

    MartinsPedro M. G. and L.Alan Winters. 2004. “When Comparative Advantage Is Not Enough: Business Costs in Small Remote Economies.”

    • Search Google Scholar
    • Export Citation

    MayerThierry and SoledadZignago. 2011. “Notes on CEPII’s Distances Measures: The GeoDist Database.” Working Paper 2011–25 Centre d’ Études Prospectives et d’Informations Internationales Paris.

    • Search Google Scholar
    • Export Citation

    ReddingStephen and AnthonyVenables. 2004. “Economic Geography and International Inequality.” Journal of International Economics62 (1): 5382.

    • Search Google Scholar
    • Export Citation

    South Pacific Applied Geoscience Commission and United Nations Environment Programme. 2005. “Building Resilience in Small Island Developing States (SIDS): The Environmental Vulnerability Index.” Paper presented at the Mauritius International MeetingJanuary1014.

    • Search Google Scholar
    • Export Citation

    World Bank. 2007. “Defining a Small Economy.”

    ZhuMin. 2012. “Sheer Distance a Factor as Pacific Islands Seek to Boost Growth.” IMF Survey MagazineMarch21.

    • Search Google Scholar
    • Export Citation

This chapter is a revised version of Becker (2012), and benefited from helpful comments by Hoseung Lee, Cynthia Rohan, John Rolle, Piers Merrick, Shanaka Peiris, Dominique Simard, Patrizia Tumbarello, and Yongzheng Yang. I thank Lucy Pan for discussions about the vulnerabilities of small island states in the Pacific while visiting Tuvalu. Data for the Environmental Vulnerability Index were gratefully received from Ursula Kaly at the South Pacific Applied Geoscience Commission.


Given the importance of remittances and transfer payments as sources of income in many of these states, GNI is generally accepted as a more appropriate measure than the value added measured by GDP. See Annex 9.2 for metadata.


Fifteen states in the raw World Bank database were excluded from the initial sample of 65 on the basis that comparable data for other variables of interest, mainly GNI, were not available. These were American Samoa, Aruba, Bermuda, Cayman Islands, Curaçao, Faeroe Islands, French Polynesia, Guam, Mayotte, New Caledonia, Northern Mariana Islands, Sint Maarten, St. Martin, Turks and Caicos Islands, and the United States Virgin Islands.


Groupings based on World Bank regional designations.


Under the IDA framework, the exceptional circumstances of Marshall Islands and Tuvalu are recognized by an exemption from the income per capita threshold that allows them access to IDA loans. The IMF made some adjustments to the Poverty Reduction and Growth Trust eligibility criteria in view of the unique challenges they face (IMF 2013a, 2013b). See also IDA (2001) and IMF (2009, 2012) for background to the decisions.


See Martins and Winters (2004) for a more detailed discussion of the economic factors that allow states such as Liechtenstein and Luxembourg to overcome the disadvantages of their smallness.


To investigate this issue more fully, factors such as different fixed costs arising from characteristics such as dispersion of the population and accessibility to service providers (for example, health care) would have to be considered.


For small states in the Pacific, the scope for greater integration with Australia and New Zealand might hold the most promise.


The main coordinating body in the region is the Pacific Islands Forum, which facilitates the Forum Economic Ministers’ Meeting. This type of cooperation has resulted in the formulation of the Pacific Plan, the Pacific Agreement on Closer Economic Relations, several agreements on trade, and discussion about action on climate change. There is also coordination of technical assistance and training through the Pacific Financial Technical Assistance Center. See Browne (2006) for a concise summary of the Pacific Plan.


Given that income thresholds are used as a proxy for welfare, it might also be that in the presence of significant vulnerabilities and risk aversion, welfare is notably lower than implied by income per capita.


This is not surprising given that if the land were very fertile and able to support a larger number of people, it would probably be more densely populated.


We acknowledge that while some states in the Northern Hemisphere may at face value appear to be quite arable, their proximity to the Arctic Circle severely curtails the ability to foster a productive agricultural sector.


Many of the most remote states, for a variety of reasons, do not export any goods. As a consequence, container ships that deliver imports on the first leg of the journey have no cargo for the return leg. This raises the cost of delivering containers. Many smaller island countries also do not have sufficient infrastructure for the larger, more efficient, container ships to dock. Furthermore, since fuel is a major part of shipping costs, imports that require crossing substantial maritime distances expose the importing country even further to fluctuations in the price of fossil fuels.


Once again caution is required when making inferences from the data. While the measure of distance used here is favored because it is simple and transparent, there are alternative ways to consider isolation. A country may be isolated not because of distance but because it is landlocked and surrounded by politically unstable neighbors that are subject to civil unrest. Financial isolation or connectivity to telecommunications might also be important variables.


On this measure even relatively heavily populated developed states can be considered remote from the rest of the world. New Zealand stands out as the single most distant state from all others in the world based on these calculations. Ranked at number 15, Australia is also very remote, but in part due to its size, resource endowments, and colonial ties to the Commonwealth, has been able to overcome this disadvantage more effectively than most small states. These data and the author’s calculations are not shown in this chapter, but are in Mayer and Zignago (2011).


There may also be advantages in dealing with some of the nonlinearities in the data by employing this ranking methodology.


There is no reason to presume each of the five indicators discussed in this chapter is of equal importance as implied by the weighting employed. However, this chapter does not estimate the relevant preference function, as this would differ by region and state.


However, there are some Pacific states that do not appear notably more exposed than those in other regions.


Other considerations worth bearing in mind are whether one would necessarily expect a relationship between income per capita and average growth outcomes, or whether the relationship is significantly more complicated than the treatment afforded it in this chapter.


Redding and Venables (2004) show that geography of access to markets is statistically significant and quantitatively important in explaining cross-state differences in income per capita.


Both of these states enjoyed extraordinarily rapid income growth due to significant oil and gas exploration projects as well as rising fuel prices during 2001–10.


Dropping the two outliers can also be justified on the grounds that they are almost 4 standard deviations away from the sample mean.


An informative contrast to the findings presented here and in the literature more generally is Easterly and Kraay (2000), where the authors find no empirical evidence of such relationships and conclude that small states should receive the same policy advice as larger countries.

    Other Resources Citing This Publication