5 Islands of Stability? Determinants of Macroeconomic Volatility in the Eastern Caribbean Currency Union

David Robinson, Paul Cashin, and Ratna Sahay
Published Date:
September 2006
  • ShareShare
Show Summary Details
Tobias Rasmussen and Guillermo Tolosa 

With their high degree of openness, dependence on tourism, and proneness to natural disasters, the countries of the Eastern Caribbean Currency Union (ECCU) are unusually exposed to external shocks. Nevertheless, the volatility of economic output in the ECCU has over past decades been markedly lower than in other high-middle-income countries. This chapter finds that this relative stability is explained by fiscal policies and international capital flows that are less procyclical than typically is the case in other developing countries. The scope for continued stability could be ending, however, as high public debt may force outcomes to become more procyclical.

The literature on macroeconomic volatility would predict high output volatility in the ECCU; however, the evidence in the region does not support this hypothesis. A number of studies—including Acemoglu and Zilbotti (1997), Easterly and Kraay (2000), Easterly, Islam, and Stiglitz (2000), and Pritchett (2000)—have sought to uncover the sources of economic volatility. Their findings suggest that output volatility is typically associated with a low level of income, lack of diversification, and openness to trade. One would then expect the ECCU countries to be extremely volatile, considering that their sum of imports and exports amounts to about 130 percent of GDP, their tourism receipts account for roughly half of total exports of goods and services, and they endure frequent devastating hurricanes. Historically, however, the ECCU economies have been remarkably stable. Indeed, the analysis finds that the standard indicators of vulnerability identified in the literature suggest a level of real GDP volatility that is about twice as high as that actually observed. The exceptionally low volatility of the ECCU would therefore seem related to factors that are unique to the region.

The low level of output volatility in the ECCU has so far received little attention. There is a large literature on the special vulnerability of small island states (Atkins, Mazzi, and Easter, 2000), but most of the studies do not address the fact that the ECCU countries fail to fit into this picture in terms of their historical output volatility. Berezin, Salehizadeh, and Santana (2002)—in one of the few studies to mention this anomaly—suggest five causes: macroeconomic stability; absence of large-scale social conflicts; a declining role of agriculture; low correlation between sectors; and stable export earnings. This chapter builds on their work by exploring the causes of low output volatility and discusses possible implications.

The region’s reliance on tourism has contributed to the relative stability, but it does not appear to be the dominating determinant. Dependence on tourism has not created volatility as one could have expected, as it is an unusually stable industry. Indeed, despite the lack of diversification, the volatility of overall exports of goods and services is lower in the ECCU than in the average developing country. In addition, private capital inflows into the ECCU countries are predominantly in the form of comparatively stable foreign direct investments. Nevertheless, the impact of these two sources of stability appears limited, as estimates suggest that private sector volatility in the ECCU is relatively high.

A key source of stability in the ECCU has been the ability to pursue countercyclical policies. Contrary to the procyclical tendencies in developing countries documented by Kaminsky, Reinhart, and Végh (2004), fiscal policy and international capital flows in the ECCU countries are found to have been only mildly procyclical or even countercyclical. The absence of this “when it rains, it pours” syndrome helps explain why the effects of high external vulnerability have been so muted. A key reason for the ECCU countries’ ability to pursue countercyclical policy is that they have had relatively easy access to capital in good and bad times. This, in turn, may be related to the exceptional stability of the common quasi-currency board arrangement, with the Eastern Caribbean dollar pegged to the U.S. dollar since 1976 and to the pound sterling before that. Such an explanation would also be consistent with recent studies by Acemoglu and others (2003) and Satyanath and Subramanian (2004), who argue that institutions are the fundamental determinants of economic outcomes, and that macroeconomic policies are symptoms rather than root causes of volatility. This may be a mixed blessing, however. If the monetary arrangement has provided easier access to capital it may also have contributed to the buildup of debt.1

While the ECCU countries so far have been islands of stability, there are worrying signs that this might be ending. Fiscal balances have deteriorated sharply in the ECCU since the mid–1990s, and public debt has risen rapidly. With public debt-to-GDP ratios now among the highest in the world, governments will not likely be able to borrow to the extent they have in the past when faced with the next downturn. Such procyclical tendencies would lead to greater volatility and have detrimental consequences. Ramey and Ramey (1995), for example, find that higher output volatility leads to significantly lower economic growth. In addition, given risk aversion and a limited capacity to insure, volatility itself is associated with a welfare cost that may be very large in developing countries (Pallage and Robe, 2003; Chapter 7 of this volume).

Stylized Facts: High Vulnerability but Low Volatility

The ECCU countries share a number of structural features that make them exceptionally vulnerable to external shocks. Part of this exposure is related to their very small size—the combined annual GDP of the ECCU-6, the six member countries that also belong to the IMF, is less than US$3 billion, and their total population is just 570,000.2

The countries also share a number of other features that add to their vulnerability (Table 5.1). The most visible vulnerability is exposure to natural disasters. As will be documented in Chapter 7, the ECCU countries are among the most frequently hit in the world by several types of natural disasters, primarily hurricanes but also earthquakes and volcanoes. Estimates of the costs of natural disasters are subject to considerable uncertainty, but available data suggest that the value of damage in the ECCU is equivalent to 2 percent of GDP per year on average. Some catastrophic events caused damage exceeding 100 percent of GDP, such as Hurricane David in Dominica (1979), Hurricane Georges in St. Kitts and Nevis (1998), and Hurricane Ivan in Grenada (2004). Over the past three decades, the 12 most damaging disasters in the region were associated with a median 2.2 percentage point same-year decline in the growth rate of real GDP, which has clearly contributed to output volatility.

Table 5.1.Selected Indicators of Exposure to Exogenous Shocks
Number of



Divided by


Average Annual

Damage from

Natural Disasters

(Percent of GDP)
Imports of

Goods and


(Percent of GDP)
Exports of

Goods and


(Percent of GDP)



of total exports)
Antigua & Barbuda7990.7S47663
St Kitts & Nevis1,1714.0754639
St Lucia4592.0715674
St Vincent & the Grenadines7511.0615342
Small island developing states3921.8705736
All countries1000.7494418
Sources: IMF, World Economic Outlook database; World Bank, World Development Indicators database; Emergency Disasters Data Base (EM-DAT) (CRED, 2005); and IMF staff estimates.Notes: The data on natural disasters refer to 1970—2004; all other data are for 2000. Figures for country groups are unweighted averages. ECCU-6 denotes the six countries in the Eastern Caribbean Currency Union that also are members of the IMF.

A second striking feature of the ECCU economies is their dependence on international trade. Openness renders countries vulnerable to volatile international markets and has been found to lead to high output volatility (Easterly and Kraay, 2000). In the ECCU countries, the sum of exports and imports is very high, at about 130 percent of GDP. Imports alone represent about 70 percent of GDP, reflecting the high dependence of the tourism sector and domestic markets on imported goods.

A third source of vulnerability is the lack of economic diversification. A concentrated production structure can be expected to lead to higher output volatility (Jansen 2004;Mobarak, 2004). In the ECCU, the large export sector is heavily dependent on tourism. In addition, a single agricultural crop typically dominates merchandise exports.

All the standard indicators of economic vulnerability would suggest that the ECCU economies are among the most vulnerable in the world.

Table 5.2.The Commonwealth Composite Vulnerability Index Rankings



to Disasters

Vulnerability Index
Antigua & Barbuda56972
St. Kitts & Nevis16596729
St. Lucia9403719
St. Vincent & the Grenadines32444324
Source: Atkins, Mazzi, and Easter (2000).Notes: The composite index is a weighted average of the three variables, where export dependence is measured by exports of goods and services as a fraction of GDP, export diversification is given by the UNCTAD diversification index for merchandise exports, and vulnerability to natural disasters is given by the percent of population affected by disasters. The weights are given by the importance of these variables in determining output volatility. The sample has 111 countries. For each measure, the country deemed the most vulnerable is assigned a ranking of “1.”

A number of studies have sought to synthesize the different variables into a composite vulnerability index, as exemplified in Table 5.2. By this measure, all of the ECCU countries are among the top 30 of the 111 countries considered, with Antigua and Barbuda taking second place.3

The ECCU countries may also lack resilience to adverse events stemming from a low capacity to absorb shocks. Hausmann and Gavin (1996) find that inflexible exchange rate regimes contribute to higher macroeconomic volatility because of their inability to absorb real shocks. Thus, most of the adjustment must take place via changes in output. From this perspective, the ECCU’s fixed exchange rate regime would be a source of added volatility. Also contributing to low resilience is that even if some relative price movements take place, the responsiveness of tourism tends to be smaller than that of other exports.4 Consequently, countries would find it difficult to expand tourism to compensate for downturns in other parts of the economy. This general tendency may well be especially pronounced in the ECCU, where the dominant form of high-end tourism is widely considered to be price inelastic.

Table 5.3.Volatility of Real GDP, 1971–2003(Annual percent change)
Number of Years with

Growth Less than:

Real GDP




−2 percentAverage


4 percent
Antigua & Barbuda4.93.10.613
St Kitts & Nevis4.82.80.613
St Lucia4.65.11.115
St Vincent & the Grenadines4.53.60.812
Small island developing states (34)
All countries (175)
Low-income (55)
Middle-income (50)
High-middle-income (33)
High-income (37)
Developing countries by region
Caribbean (14)
Latin America & the Caribbean (31)
East Asia & Pacific (17)
South Asia (7)
Europe & Central Asia (26)
Sub-Saharan Africa (44)
Middle East & North Africa (13)
Sources: IMF, International Financial Statistics and World Economic Outlook databases; Eastern Caribbean Currency Union country authorities; and IMF staff estimates.Notes: Figures for country groups are simple averages, with the number of countries in parentheses. ECCU-6 denotes the six countries in the Eastern Caribbean Currency Union that also are members of the IMF.

Economic Volatility: Cross-Country Evidence

Surprisingly, the ECCU countries have been islands of stability in a volatile developing world. In other words, the high degree of vulnerability identified in the previous section has not been associated with the high output volatility one would have expected. It is well documented that output volatility tends to be markedly higher in developing than in high-income countries (Agénor, McDermott, and Prasad, 2000) and that small economies have experienced higher volatility than large economies (Easterly and Kraay, 2000). However, as Table 5.3 indicates, the volatility of real GDP growth in the ECCU—as measured by the standard deviation over the past three decades—has on average been lower than in virtually Surprisingly, the ECCU countries have been islands of stability in a volatile developing world. In other words, the high degree of vulnerability identified in the previous section has not been associated with the high output volatility one would have expected. It is well documented that output volatility tends to be markedly higher in developing than in high-income countries (Agénor, McDermott, and Prasad, 2000) and that small economies have experienced higher volatility than large economies (Easterly and Kraay, 2000). However, as Table 5.3 indicates, the volatility of real GDP growth in the ECCU—as measured by the st and ard deviation over the past three decades—has on average been lower than in virtually all regions of the world. This finding is robust to alternative measures of volatility—for example, measured by the coefficient of variation (the standard deviation divided by the mean) or the fraction of years with growth below a certain threshold. Moreover, volatility in the ECCU has been on a declining trend since the mid–1980s (Figure 5.1). This section seeks to explain the reason for this remarkably low level of output volatility 5

Figure 5.1.Eastern Caribbean Currency Union: Real GDP

Source: Eastern Caribbean Currency Union country authorities.

The core variables that have been used to explain output volatility in cross-country regressions cannot account for the unusual stability in the ECCU. Table 5.4 shows the results of regressing countries’ historical output volatility on a series of explanatory variables capturing potentially relevant characteristics of the economies, including measures of vulnerability discussed in the previous section. Regression (1) includes a core set of explanatory variables that have emerged in the literature on determinants of volatility: institutional quality; the degree of openness; the size of the economy; and average per capita income.6 It shows significant coefficients associated with each of these variables, with the signs of the first three as expected. Contrary to some previous findings (such as Jansen, 2004) and the simple correlation identified in Table 5.3, the coefficient on per capita income is positive, suggesting that higher income does not lower volatility when controlling for the other variables. While these variables explain a sizable share of the cross-country variation in output volatility, the regression fails to explain the ECCU’s very low volatility, as reflected by the negative and highly significant coefficient for the ECCU dummy variable. Indeed, the magnitude of the coefficient on the ECCU dummy suggests that the region’s volatility has been only half of what the other variables imply. In contrast, the coefficient on the dummy variable for other small island developing states is much smaller and barely significant at the 10 percent level.

Table 5.4.Determinants of Output Volatility, 1971–2003(Cross-country OLS regressions with the natural logarithm of the standard deviation of annual real GDP growth as the dependent variable)
Dummy for
ECCU-6−0.506 (0.202)**−0.498(0.229)**−0.469 (0.214)**−0.666(0.243)***−0.385 (10.210)*−0.093 (0.241)
Other small island developing states−0.177 (0.105)*−0.051 (0.148)−0.077 (0.108)−0.064 (0.174)−0.063 (1.0.122)−0.002 (0.165)
Institutions (regulatory quality)−0.297 (0.050)***−0.314 (0.071)***−0.230 (0.055)***−0.252 (0.079)***−0.276 (0.072)***−0.263 (0.079)**
Openness (ratio of exports to GDP)0.002 (0.001)*0.000 (0.002)0.002 (0.001)*0.000 (0.003)0.001 (0.002)0.007 (0.003)**
Size (U.S. dollar GDP+)−0.085 (0.020)***−0.079 (0.028)**−0.083 (0.021)***−0.074(0.031)**−0.079 (0.023)**−0.025 (0.028)
Income (PPP GDP per capita in 1970+)0.131 (0.041)***−0.026 (0.072)0.085 (0.049)*0.011 (0.065)0.095 (0.057)*0.034 (0.061)
Terms of trade volatility0.000 (0.002)
Export concentration0.157(0.223)
Agriculture share in GDP−0.007 (0.006)
Damage from natural disasters(0.00l)
Dummy tor developing countries in
Latin America & Caribbean0.040 (0.124)
East Asia & Pacific0.010 (0.173)
South Asia0.291 (0.205)
Europe & Central Asia0.363 (0.133)***
Sub-Saharan Africa0.068 (0.162)
Middle East & North Africa0.292 (0.159);*
M2 ratio to GDP0.005 (0.002)**
Financial flows ratio to GDP−0.006 (0.012)
standard deviation0.019(0.006)***
Exchange rate regime−0.006 (0.051)
Transfers ratio to GDP−0.002 (0.010)
Services share in exports−0.002 (0.003)
Foreign direct investment ratio to GDP0.008 (0.005)
Public consumption ratio to GDP0.010(0.010)
standard deviation0.000 (0.011)
Fiscal procyclicality0.037 (0.010)**
Constant0.842 (0.260)***2.054 (0.600)***1.045 (0.397)***1.347 (0.396)***1.074 i0.374j***0.757 (0.357)**
Number of observations1681071688712789
Sources: See Appendix 5.1.Notes: Figures in parentheses are standard deviations, and *, **, *** indicate significance at, respectively, the 10, 5, and 1 percent level. Variables denoted with a † are expressed in natural logarithms. PPP denotes purchasing power parity. OLS denotes ordinary least squares.

Other explanatory variables used in previous studies do not explain the puzzle. Regression (2) includes measures of terms of trade volatility, export concentration, exposure to natural disasters, and the importance of agriculture in the economy. None of these variables has significant coefficients, although the explanatory power of the regression increases and some of the previous coefficients lose their significance.7 Regression (3) introduces a series of regional dummies. Here the dummies for developing countries in Europe and Central Asia and the Middle East and North Africa are found to be statistically significant, reflecting those countries’ high volatility, while the other coefficients are broadly similar to those in regression (1). In both regressions the coefficient on the ECCU dummy variable remains strongly negative. This shows that even a broad set of explanatory variables fail to account for the stability of the ECCU output, with regressions (1) to (3) all pointing to predicted volatility about double the actual level.

Table 5.5.Selected Indicators

Deviation of

U.S. Dollar




per year)


in Total






of GDP)




of GDP)

Money (M2)


of GDP)
Antigua & Barbuda36753.95.367
St. Kitts & Nevis21478.116.865
St. Lucia17572.67.552
St. Vincent & the Grenadines13524.315.452
Small island developing states23453.96.252
All countries22262.44.643
Sources: IMF, World Economic Outlook database.Notes: The data refer to 1971–2003, except for current transfers and foreign direct investment, which are averages over 1995–2003. ECCU-6 denotes the six countries in the Eastern Caribbean Currency Union that also are members of the IMF.

Several other factors could explain the relative stability of the ECCU. Part of the explanation may be that the volatility of exports of goods and services has not been particularly high given the lack of diversification (Table 5.5). This can be attributed to the relative stability of the large tourism industry and, to a lesser extent, the stable export prices afforded to agricultural exports under preferential trading arrangements. Another possible source of stability is that the ECCU nations have a large diaspora and receive substantial remittances (see Chapter 9) that could potentially help offset economic difficulties, although Cashin and Wang (2005) find that such transfers have tended to be procyclical. In addition, the countries receive exceptionally high levels of foreign direct investment (FDI) that are relatively stable in comparison to inflows of portfolio investment. Finally, the ECCU countries are highly monetized compared with other developing countries, which may have helped buffer adverse shocks.

Accounting for the factors noted in the previous paragraph still leaves unresolved questions. Regression (4) introduces a series of variables related to the financial sector, and shows the M2-to-GDP ratio and the standard deviation of international financial flows entering with significant positive coefficients (Table 5.4). However, after controlling for these factors, the coefficient on the ECCU dummy becomes even more negative. Regression (5) includes measures of the magnitude of service exports, international transfer receipts, and FDI to capture some of the other atypical features of ECCU economies. None of these variables has a significant impact on output volatility, although the coefficient on the ECCU dummy is slightly reduced.

Table 5.6.Indicators of Economic Volatility, 1984–2003(Standard deviations)

GDP Growth

(Percent per year)


(Percent of GDP)
“Private” Real

GDP Growth

(Percent per year)1
Antigua & Barbuda3.53.44.9
St. Kitts & Nevis3.08.09.8
St. Lucia5.02.56.6
St. Vincent & the Grenadines3.82.56.3
Small island developing states4.73.96.8
Low-income countries5.44.06.5
Low-middle-income countries5.33.26.3
High-middle-income countries5.64.06.7
High-income countries3.32.34.2
Sources: IMF, International Financial Statistics and World Economic Outlook databases; Eastern Caribbean Currency Union country authorities; and IMF staff estimates.Note: ECCU-6 denotes the six countries in the Eastern Caribbean Currency Union that also are members of the IMF.

The key reason for the relatively low volatility of output in the ECCU appears to have been the countercyclical fiscal policy pursued by national governments. Excluding the public sector’s contribution to GDP suggests that private sector output volatility has been relatively high in the ECCU (Table 5.6).8 The high level of volatility in the private sector is what one would expect given the high level of vulnerability to external shocks identified above. This suggests that the low level of overall volatility is a result of developments in the public sector.

Table 5.7.Measures of Procyclicality
Fiscal PolicyCapital Flows
ECCU-6 average1.5−1.4
Antigua & Barbuda7.06.2
St. Kitts & Nevis1.8−2.3
St. Lucia2.50.6
St. Vincent & the Grenadines2.2−6.0
Small island developing states4.2. . .
Low-income countries8.60.3
Middle-low-income countries4.21.2
Middle-high-income countries6.51.4
High-income countries0.40.1
Sources: IMF staff estimates for Eastern Caribbean Currency Union countries, 1983–2004; Kaminsky, Reinhart, and Végh (2004) for 104 other countries, 1960–2003.Notes: Fiscal policy refers to the average annual growth in central government real expenditure. Capital flows refer the financial account balance in percent of GDP. In both cases, the index is computed as the difference between the average value in good times (real GDP growth above median) and the average value in bad times (real GDP growth below median). Higher values are thus associated with more procyclical developments. ECCU-6 denotes the six countries in the Eastern Caribbean Currency Union that also are members of the IMF.

Fiscal policy in the ECCU has been relatively countercyclical. Government expenditure in developing countries tends to grow much faster in good times than in bad times (Table 5.7). Except for Antigua and Barbuda, this tendency is mostly absent in the ECCU. Indeed, by this measure, fiscal policy in the ECCU has on average been almost as countercyclical as in high-income countries.9 The critical importance of fiscal policy is evident in regression (6), where the measure of fiscal procyclicality enters with a positive and strongly significant coefficient (Table 5.4).10 This regression has the highest R-squared of the six, with the measures of institutional quality and openness being the only other variables associated with significant coefficients. Moreover, in this regression the coefficient on the ECCU dummy variable is substantially reduced and becomes statistically insignificant.

An even more striking result is that international capital flows have been much more countercyclical in the ECCU countries than in other regions, including the high-income countries. Since public sector borrowing has driven a large part of capital inflows into the ECCU, the countercyclicality of international capital flows is to some extent the result of the low degree of procyclicality in fiscal policy. Both measures reflect that the ECCU governments have had unusually easy access to credit, allowing them to borrow in periods when other developing countries have typically been cut off. Importantly, Antigua and Barbuda, the one ECCU country that has been decidedly procyclical, is also the one that has been cut off from traditional international capital markets following defaults on its external debt dating back more than a decade (Table 5.7).

The quasi-currency board arrangement is likely to have been an important contributor to output stability in the ECCU. Given that a countercyclical stance is a goal that governments typically aspire to but are often unable to achieve, the question arises as to why the ECCU countries have been less financially constrained than other developing countries. Although the cross-country regressions do not point to any significant effect from the exchange rate regime, the relatively countercyclical nature of fiscal policy and international capital flows in the ECCU could be related to the stability provided by the monetary arrangement.11 Having maintained a fixed exchange rate against the U.S. dollar for almost three decades, the system has undoubtedly contributed to keeping inflation and interest rates low and stable, and has facilitated the development of deep financial systems. Indeed, there are very few other countries in the world that have maintained a fixed exchange rate for so long. (Panama is the only other country with an equally impressive record in the Reinhart-Rogoff [2002] dataset.)12 That the cross-country regressions do not detect a significant impact of the exchange arrangement may simply reflect that there are very few other countries that have managed to establish such enduring pegs. Also, while a fixed exchange rate may help provide access to credit and thereby facilitate the operation of countercyclical forces, the overall impact on output volatility is ambiguous, as the effect of reduced price flexibility identified by Hausmann and Gavin (1996) would work in the opposite direction. Another possible explanation is that, with the exception of Antigua and Barbuda, the ECCU countries by and large have had (until recently) an excellent record of remaining current on sovereign borrowings, a factor that has been found to be an important determinant of a country’s borrowing capacity (Reinhart, Rogoff, and Savastano, 2003).

Concluding Remarks

Despite the frequency of real shocks, the volatility of output in ECCU countries has been surprisingly low. This exceptionally low volatility has been associated with the fact that fiscal policy has been markedly less procyclical than in other developing countries. The ability to borrow as a result of a good record of debt repayment in most countries, and the stability of the monetary system, are two important factors that have allowed ECCU countries to pursue countercyclical policies.

Cross-country experience shows that countercyclical fiscal policy is one of the main drivers of low economic volatility. The analysis indicates that the cyclicality of government expenditure is a key determinant of output volatility. High output volatility is also strongly linked to low institutional quality, and there is partial evidence of a positive impact from openness, small size of the economy, high per capita income, and a high degree of monetization. Several of these variables may be interrelated and it is therefore difficult to pinpoint their individual significance. It is clear, however, that the countercyclical fiscal policy pursued by ECCU countries has helped dampen what would otherwise have been a much higher level of volatility.

Future stability of the ECCU economies will depend on the continued capacity to pursue fiscal policies that are markedly less procyclical than in other developing countries. If public debt continues to rise and borrowing limits are reached, expenditure reductions in downturns may become inevitable, thereby contributing to greater output volatility. Debt distress is already evident in Antigua and Barbuda, Dominica, and Grenada, indicating that it will be difficult to pursue expansionary fiscal policies in the near future.

Appendix 5.1. Data Sources
Real GDPCountry authorities for data on the ECCU. Otherwise, International Financial Statistics(series …99BVPZF…) where available, and else World Economic Outlook (WEO) database (seriesW…NGDP_R).
SizeGross domestic product, current prices, U.S. dollars from WEO (series W…NGDPD). Values used are the average levels over 1971–2003.
IncomePurchasing power parity (PPP) per capita from WEO (series W…PPPPC). Values used are the levels in 1970.
InstitutionsRegulatory quality from the World Bank’s World-wide Governance Research Indicators dataset. Scores range between –2.5 (worst) and +2.5 (best). Values used are average scores for 1996–2002.
OpennessRatio of exports of goods and services to GDP, both in U.S. dollars, from WEO (series W…TX and W…NGDPD). Values used are the average levels over 1971–2003.
Terms of trade

Terms of trade, goods and services index from WEO (series W…TT). Values used are the standard deviations over 1971–2003.
Export concentrationConcentration of exports (goods only) index from UN Conference on Trade and Development (UNCTAD). Values used are the levels in 2000.
Agriculture share

in GDP
Agriculture value added (percent of GDP) from the World Bank’s World Development Indicators (series …NVAGRTOTLZS). Values used are the average levels over 1971–2003, including countries with partial series if there are at least 10 observations.
Damage from

natural disasters
Cumulative damage during 1970–2002 (in percent of GDP) from Chapter 7, based on EM-DAT data.
M2 ratio to GDPBroad money and GDP, both in national currency, from WEO (series W…FMB and W…NGDP). Values used are the average levels over 1971–2003.
Financial flows

ratio to GDP
Financial account balance and GDP, both in U.S. dollars, from WEO (series W…BF and W… NGDPD). Values used are the average levels over 1971–2003, and the standard deviation over the same period.
Exchange rate

Coarse annual classification by Reinhart and Rogoff (2002) available at Values used are average levels over 1970–2002.
Transfers ratio

to GDP
Current private transfers and GDP, both in U.S. dollars, from WEO (series W…BTRP and W… NGDPD). Values used are the average levels over 1995–2003.
Services share

in exports
100 minus export of goods in percent of value of exports of goods and services, both in U.S. dollars, from WEO (series W…BXG and W…TX). Values used are average levels over 1971–2003.
FDI ratio to GDPDirect investment in reporting economy in percent of GDP, both in U.S. dollars, from WEO (series W…BFDI and W…NGDPD). Values used are the average levels over 1995–2003.
Public consumption

ratio to GDP
Public consumption expenditure in percent of GDP, both in national currency, from WEO (series W…NCG and W…NGDP). Values used are the average levels over 1971–2003, and the st and ard deviation over the same period.
Fiscal procyclicalityFrom Kaminisky, Reinhart, and Végh (2004), augmented with own calculation for the ECCU countries using the same methodology based on data from ECCU country authorities.

Income classification according to World Bank (; regional classification according to World Bank (; and small island developing states according to Small Island Developing States Network (

    Acemoglu, D., and F.Zilbotti,1997, “Was Prometheus Unbound by Chance? Risk, Diversification, and Growth,”The Journal of Political Economy,Vol. 105,No. 4,pp. 709–51.

    Acemoglu, D., S.Johnson, J.Robinson, and Y.Thaisharown,2003, “Institutional Causes, Macroeconomic Symptoms: Volatility, Crises and Growth,”Journal of Monetary Economics,Vol. 50,No. 1,pp. 49–123.

    Agénor, P., C.J.McDermott, and E.Prasad,2000, “Macroeconomic Fluctuations in Developing Countries: Some Stylized Facts,”The World Bank Economic Review,Vol. 14,No. 2,pp. 251–85.

    Atkins, J.P., S.Mazzi, and C.D.Easter,2000, A Commonwealth Vulnerability Index for Developing Countries: The Position of Small States(London: Commonwealth Secretariat).

    Berezin, P., A.Salehizadeh, and E.Santana,2002, “The Challenge of Diversification in the Caribbean,”IMF Working Paper 02/196(Washington: International Monetary Fund).

    Briguglio, L.,1992, Preliminary Study on the Construction of an Index for Ranking Countries According to their Economic Vulnerability(Geneva: UN Conference on Trade and Development).

    Briguglio, L.,1993, “The Economic Vulnerabilities of Small Island Developing States,”study commissioned by CARICOM for the Regional Technical Meeting of the Global Conference on the Sustainable Development of Small Island Developing States, Port of Spain, Trinidad and Tobago,July.

    Briguglio, L.,1995, “Small Island States and their Economic Vulnerabilities,”World Development,Vol. 23,No. 9,pp. 1615–32.

    Briguglio, L.,1997, “Alternative Economic Vulnerability Indices for Developing Countries,”report prepared for the Expert Group on Vulnerability Index, United Nations Department of Economic and Social Affairs (DESA),December.

    Cashin, P., and P.Wang,2005, “Macroeconomic Fluctuations in the Eastern Caribbean Currency Union,” in Eastern Caribbean Currency Union: Selected Issues, IMF Country Report No. 05/305(Washington: International Monetary Fund),pp. 3–17.

    Centre for Research on the Epidemiology of Disasters (CRED), 2005, Emergency Disasters Data Base (EM-DAT), Brussels.Available via the Internet:

    Crowards, T.,2000, “An Index of Inherent Economic Vulnerability of Developing Countries,”Staff Working Paper No. 6/00(Bridgetown, Barbados: Caribbean Development Bank).

    Crowards, T., and W.Coulter,1998, “Economic Vulnerability in the Developing World with Special Reference to the Caribbean,”Working Paper(Bridgetown,Barbados: Caribbean Development Bank).

    Easterly, W., and A.Kraay,2000, “Small States, Small Problems? Income, Growth, and Volatility in Small States,”World Development,Vol. 28,No. 1,pp. 2013–27.

    Easterly, W.,R.Islam, and J.Stiglitz,2000, “Shaken and Stirred: Explaining Growth Volatility,” inAnnual Bank Conference on Development Economics,ed. byB.Pleskovic and N.Stern(Washington: World Bank)

    Fiaschi, D., and A.Lavezzi,2003, “On the Determinants of Growth Volatility: a Nonparametric Approach,”University of Pisa Discussion Paper No. 25(Pisa, Italy: University of Pisa).

    Hausmann, R., and M.Gavin,1996, “Securing Stability and Growth in a Shock Prone Region: The Policy Challenge for Latin America,”IADB Working Paper No. 315(Washington: Inter-American Development Bank).

    Jansen, M.,2004, Income Volatility in Small and Developing Economies: Export Concentration Matters(Geneva: World Trade Organization).

    KaminskyG., C.Reinhart, and C.Végh,2004, “When it Rains, it Pours: Procyclical Capital Flows and Macroeconomic Policies,”NBER Working Paper No. 10780(Cambridge, MA: National Bureau of Economic Research).

    Mobarak, A.,2004, “Determinants of Volatility and Implications for Economic Development,”Political and Economic Change Working Paper PEC 2004–0001, (Boulder: University of Colorado).

    Pain, N., and D.Van Welsum,2004, “International Production Relocation and Exports of Services,”OECD Economic StudiesNo. 38,pp. 67–94.

    Pallage, S., and M.Robe,2003, “On the Welfare Cost of Economic Fluctuations in Developing Countries,”International Economic Review,Vol. 44,No. 2,pp. 677–98.

    Pritchett, L.,2000, “Understanding Patterns of Economic Growth: Searching for Hills Among Plateaus, Mountains, and Plains,”World Bank Economic Review,Vol. 14,No. 2,pp. 221–50.

    Ramey, G., and V.Ramey,1995, “Cross-Country Evidence on the Link between Volatility and Growth,”American Economic Review,Vol. 85,No. 5,pp. 1138–51.

    Reinhart, C., and K. Rogoff,2002, “The Modern History of Exchange Rate Arrangements: A Reinterpretation,”NBER Working Paper No. 8963(Cambridge, MA: National Bureau of Economic Research).

    Reinhart, C., K. Rogoff, and M.Savastano,2003, “Debt Intolerance,”NBER Working Paper No. 9908(Cambridge, MA: National Bureau for Economic Research).

    Satyanath, S., and A.Subramanian,2004, “What Determines Long-Run Macroeconomic Stability? Democratic Institutions,”IMF Working Paper 04/215(Washington: International Monetary Fund).

    World Bank, 2003, “Caribbean Economic Overview 2002: Macroeconomic Volatility, Household Vulnerability, and Institutional Responses, Report” No 24165-LAC(Washington: World Bank).

See Chapter 3 for a discussion of the moral hazard problem inherent in the monetary arrangement.

The ECCU-6 are Antigua and Barbuda, Dominica, Grenada, St. Lucia, St. Kitts and Nevis, and St. Vincent and Grenadines.

Other vulnerability indexes have been produced by Briguglio (1992, 1993, 1995, 1997), Crowards and Coulter (1998), and Crowards (2000). These other indexes also consider transport costs and dependence on strategic imports as sources of vulnerability. The general conclusion is that small island states tend to be more economically vulnerable than other groups of countries.

See Pain and Van Welsum (2004) and references therein for a discussion on the relatively low price elasticity of tourism and other services in comparison with merchandise trade.

See World Bank (2003) for additional background on growth and volatility in the ECCU. The study presents data showing that consumption volatility in the region has been higher than output volatility and also relatively high in comparison with other countries. However, data on consumption levels in the ECCU are weak, and the high degree of consumption volatility may reflect a measurement problem.

See Appendix 5.1 for a detailed description of the data sources and definitions. The measure of institutional quality is an index of regulatory quality developed by the World Bank that captures the incidence of market-unfriendly policies and perceptions of the burdens imposed by excessive regulation.

Others have found these variables to have a significant impact on volatility. For example, Atkins, Mazzi, and Easter (2000) find a positive impact from susceptibility to natural disasters and export concentration; Easterly and Kraay (2000) find a positive impact from terms of trade volatility; and Fiaschi and Lavezzi (2003) find a negative impact from the size of the agricultural sector. Some of these differences in results may reflect different estimation periods and data sources, but they may also reflect that those studies fail to include important control variables.

Fiscal data for the ECCU is only available since 1983. However, the pattern of output volatility is broadly the same in the post–1983 period as in the 1971–2003 period considered earlier.

The applied measures of cyclicality are from Kaminsky, Reinhart, and Végh (2004).

There is a possibility that fiscal policy procyclicality, at least in part, depends on output volatility rather than the other way around, which would impair the statistical properties of the regression. Nevertheless, it seems likely that fiscal policy would depend more on the root causes of volatility, such as the size and openness of the economy. Also, similar endogeneity issues present themselves with respect to several of the other variables, notably the measure of institutional quality. Other studies have attempted to correct for potential endogeneity problems by using instrumental variable techniques (Acemoglu and others, 2003; Satyanath and Subramanian, 2004), but they still find that institutions have an important impact on volatility.

The Reinhart and Rogoff (2002) measure of exchange rate flexibility takes values on a scale of 1 to 15, with 1 indicating regimes with no separate legal tender. The ECCU countries all receive 2s (for pre-announced peg or currency board arrangement) throughout the 1940–2001 sample period. See Appendix 5.1 for additional details.

Among IMF member countries, there are seven countries aside from the ECCU countries that have managed to maintain an exchange rate relative to the U.S. dollar that has never varied by more than 1 percent since 1980 on a year-average basis: The Bahamas, Bahrain, Barbados, Belize, Djibouti, Panama, and Qatar. However, as with the Reinhart and Rogoff (2002) index, a dummy variable for these countries is not associated with a significant coefficient in the cross-country-regression analysis.

    Other Resources Citing This Publication