Eastern Caribbean Currency Union: Selected Issues
Author: International Monetary Fund
Publication Date:
02 Jun 2009
eISBN:
9781451811742
Language:
English
Keywords:
ISCR; CR; ECCU economy; ECCU country; World GDP growth; ECCU region; multi-country risk pool; extent Grenada; outcomes of debt ratios; Climate change; Debt sustainability
Download PDF (487 KB)

The Eastern Caribbean Currency Union (ECCU) countries’ economies are heavily dependent on the United States for foreign direct investment, mainly in the tourism sector. The Selected Issues paper discusses economic development and policies of the ECCU. About one-third of the stayover tourists to the ECCU countries are from the United States., the top tourist-source country. The flow of remittances is also an important channel of influence, reflecting the significant proportion of Caribbean migrants living in the United States.

Abstract

The Eastern Caribbean Currency Union (ECCU) countries’ economies are heavily dependent on the United States for foreign direct investment, mainly in the tourism sector. The Selected Issues paper discusses economic development and policies of the ECCU. About one-third of the stayover tourists to the ECCU countries are from the United States., the top tourist-source country. The flow of remittances is also an important channel of influence, reflecting the significant proportion of Caribbean migrants living in the United States.

VI. Insuring Against Natural Disasters in the Caribbean11

A. Introduction

1. This paper examines the vulnerability of public finances in the Caribbean to the occurrence of natural disasters (in particular, hurricanes), and illustrates how catastrophic risk insurance could significantly improve public debt sustainability through optimal insurance coverage.2 We apply the model used in Borensztein, Cavallo, and Valenzuela (2008) to fiscal debt sustainability analysis for the six highly-indebted Fund member countries of the ECCU.3

B. Debt Dynamics and Natural Disasters in the ECCU

2. The origin of unsustainable public debt in the ECCU countries has been attributed to a combination of exogenous shocks and policy slippages. Previous studies such as Sahay (2006) concluded that the rapid increase in debt reflected fiscal expansion related to both policy slippages and insufficient fiscal planning for anticipated (e.g., decline in preferential access) and unanticipated (e.g., reconstruction costs after natural disasters) adverse shocks. With the ECCU countries among the world's most vulnerable to natural disasters (Table 1), Rasmussen (2006) found that natural disasters have had a pronounced macroeconomic impact on these countries' fiscal and external balances, suggesting an important role for preventive measures.

Table 1.

ECCU: Exogeneous Shocks and Economic Policy Outcomes, 1988–2007

article image
Sources: Country authorities' data; IMF, World Economic Outlook; EM-DAT, Emergency Events Database; and authors' calculations.

3. Small island economies face larger constraints in absorbing the fiscal impact of natural disasters. The limited size of these economies prevents sufficient economic diversification that could help to mitigate economic losses and facilitate the recovery process. More importantly, high public indebtedness has left little or no fiscal room for maneuver in the event of an adverse economic shock. At the same time, high indebtedness often limits further access to borrowing or implies significantly higher borrowing costs to meet reconstruction expenses after natural disasters.

4. Reliance on donor assistance in the aftermath of a natural disaster is often problematic, with aid transfers being too late and too little. Table 2 shows the aggregate change in transfers from donors in the year (and two subsequent years) that the natural disaster occurs. The data suggest that while overall average transfers per hurricane year are higher than in an average year, transfers, in many cases, failed to pick up promptly in the year or immediate year following the occurrence of natural disasters, thus reducing the relief impact of such assistance in meeting liquidity gaps.

Table 2.

Transfer Inflows in the Aftermath of Natural Disasters, 1970–2008

(In percent of GDP)

article image
Source: IMF, World Economic Outlook; Araujo (forthcoming, 2009); and authors' calculations.Note: Years in bold denote the existence of either storms or floods. The transfer index is defined as:
tr0dy−1=tr0−tr−1y−1+tr1−tr−1y−1+tr2−tr−1y−1
where tr td denotes the total accumulated transfers due to the disaster in period t, tr t denotes the actual transfer that occurred in period t and y t denotes GDP in period t.

C. Improving Debt Sustainability with Disaster Insurance

5. Historical data indicate that the probability of occurrence of a hurricane of any category in a given year is about 18 percent for the ECCU region. For each of the ECCU countries, Figure 1 shows the probable maximum loss from hurricanes for return periods of one in 18, 20, 30, 50, 100, 200, 250 and 500 years, respectively, as estimated by the World Bank (2006). 4 5 As summarized in Table 3, historical data would imply that Dominica is the most vulnerable while Grenada is the least affected by hurricanes among the ECCU countries.6 The estimated fiscal loss caused by a hurricane with a return period of 1-in-30 years would be the highest in Dominica at 11 percent of GDP and the lowest in Grenada at less than 1 percent of GDP.

Figure 1:
Figure 1:

ECCU: Estimated Cost of Hurricanes Probable Maximum Loss

(in million U.S. dollars) by Return Period (years)

Citation: IMF Staff Country Reports 2009, 176; 10.5089/9781451811742.002.A006

Sources: World Bank (2006); and authors' calculations.
Table 3.

Probability Distribution and Cost of Hurricanes

article image
Sources: World Bank (2006), Table A5.2; and authors' calculations.

The return period is an estimate of the interval of time between given disaster events of a certain intensity or size.

6. In the event of a natural disaster, revenue losses and expenditure increases due to reconstruction needs could easily lead to explosive debt paths given the already high indebtedness of ECCU countries. Central government debt-to-GDP ratios ranged from about 55–70 percent of GDP in St. Vincent and the Grenadines, St. Lucia, and Dominica, close to 100 percent of GDP in Grenada and Antigua and Barbuda, and about 140 percent of GDP in the case of St. Kitts and Nevis in 2007. Moreover, taking into account government-guaranteed debt of public enterprises, total public debt-to-GDP ratios would be higher by about 10–40 percent of GDP.

7. ECCU governments' capacity to pursue specific fiscal targets over the medium-to long-term is subject to significant risks, particularly in light of the vulnerability to natural disasters. This can be shown using the standard debt equation, modified with a term to account for the impact of a hurricane. As shown in equation (1), the debt-to-GDP ratio (Dt) at time t is the product of the increase in the interest rate (rt) over real GDP growth (yt) and the debt-to-GDP ratio of the previous period, reduced by the fiscal primary surplus-to-GDP (pst), but increased by the fiscal cost of a hurricane, H (category) t (as a percent of GDP):

Dt=[(1+rt)/(1+yt)]Dt-1-pst+H(category)t(1)

8. To illustrate the risks, we generate a complete distribution of probable outcomes of debt ratios arising from random shocks to the economy for each ECCU member country. This is done by generating 1000 random outcomes for the stochastic variables rt, yt, pst, and Ht for each year from 2008–2020, using projections for 2008–13 as the baseline values.7 The occurrence of a hurricane for a particular year would increase the debt ratio by the estimated cost of the hurricane of the respective intensity, based on the data in Table 3. The results are shown in Figure 2, indicating the range of debt-to-GDP ratios with a 98 percent probability of occurrence. For example, in the case of St. Lucia, the projected baseline central government debt ratio without hurricanes could be between 73 to 128 percent of GDP by 2020, with the median at 99 percent. When the impact of hurricanes is taken into consideration, however, this range is projected to widen to 78–168 percent of GDP.

Figure 2.
Figure 2.

Debt Sustainability Analysis: Central Government Debt to GDP

(Confidence intervals at 98%)

Citation: IMF Staff Country Reports 2009, 176; 10.5089/9781451811742.002.A006

Source: Authors' calculations.
Figure 2a.
Figure 2a.

Debt Sustainability Analysis: Central Government Debt to GDP

(Confidence intervals at 98%)

Citation: IMF Staff Country Reports 2009, 176; 10.5089/9781451811742.002.A006

Source: Authors' calculations.
Figure 3.
Figure 3.

Debt Sustainability Analysis: Central Government Debt to GDP– Impact of Climate Change on Dominica and Grenada 1/

(Confidence intervals at 98%)

Citation: IMF Staff Country Reports 2009, 176; 10.5089/9781451811742.002.A006

Source: Authors' calculations.1/ Arrows represent change in debt trajectory with optimum insurance.

9. Reflecting the vulnerability to natural disasters, the public debt paths in ECCU countries are greatly affected by the incidence of unanticipated fiscal losses from additional expenditures and reductions in tax revenue. Our analysis also illustrates that, for Dominica and to a lesser extent Grenada, the occurrence of natural disasters could potentially reverse the direction of a declining debt trajectory.

10. Insurance coverage against natural disasters could provide some fiscal space for governments to deal with the consequences and help to reduce the associated accumulation of public debt. We consider parametric risk insurance that could be purchased on an individual country basis. The net effect on the debt-to-GDP ratio of the government's purchase of catastrophic risk insurance is modeled by adding to equation (1) the insurance premium as a percent of GDP, P (M) t, and the insurance payout as a percent of GDP, I (M, category) t:

Dt=[(1+rt)/(1+yt)]Dt-1-pst+H(category)t+P(M)t-I(M,category)t(2)

In the event that a hurricane of a specific intensity occurs and triggers the specific preset insurance payout, the insured government would receive an amount that is equal to or less than the hurricane costs depending on the level of insurance coverage.8 In our calculations, we assume a minimum annual insurance premium of 0.23 percent of GDP corresponding to an annual insurance coverage limit equivalent to 2.5 percent of GDP for all the ECCU countries.9 For instance, in the case of St. Kitts and Nevis, the occurrence of a 1-in-30 year event, corresponding to an estimated cost of 9.7 percent of GDP, would trigger an insurance payout equivalent to 2.5 percent of GDP if the government pays an insurance premium of 0.23 percent of GDP; a payout amount equivalent to 7.5 percent of GDP if the government pays a premium of 0.69 percent of GDP; or the maximum payout limit of 9.7 percent of GDP corresponding to a premium in the amount of 0.92 percent of GDP.

11. Catastrophic insurance at optimum coverage could help to improve debt sustainability. The results of our simulations are summarized in Figure 2 and Table 4. In the case of Dominica, which is the country most exposed to hurricanes, the maximum level of the debt-to-GDP ratio in 2020 is projected to increase from 69 percent in the absence of hurricanes to 150 percent, while optimum insurance coverage limits of 20 percent of GDP per annum (see last column of Table 4) could reduce the maximum level of the debt ratio to about 114 percent. The optimum insurance coverage is the point at which the benefits from a higher level of insurance coverage in terms of reducing the debt ratio are exactly offset by the cost of higher premiums. The optimum insurance coverage ranges from 5 percent of GDP for Grenada, 10 percent of GDP for St. Lucia, 12.5 percent of GDP for St. Kitts and Nevis and Antigua and Barbuda, 15 percent of GDP for St. Vincent and the Grenadines, and 20 percent of GDP per annum for Dominica (see Figure 2). Table 6 summarizes the optimum insurance coverage for each of the ECCU countries. The table also indicates the associated premium, based on commercial hurricane insurance coverage.

Table 4.

Central Government Debt in 2020

(In percent of GDP)

article image
Source: Authors' calculations.
Table 5.

Probability Distribution and Cost of Hurricanes: Simulation for Dominica and Grenada

article image
Source: Authors' calculations.

The return period is an estimate of the interval of time between given disaster events of a certain intensity or size.

Note: ATG denotes Antigua and Barbuda, DMA denotes Dominica, and KNA denotes St. Kitts and Nevis.
Table 6.

Estimated Optimum Insurance Coverage and the CCRIF

article image
Sources: World Bank (2007), Table A7.1-A7.2; and authors' calculations.

Rates-of-line denote the ratios of the premia to the maximum payouts.

Based on hypothetical individual commercial insurance premium and coverage.

12. The Caribbean Catastrophe Risk Insurance Facility (CCRIF) is a recent development which significantly improves Caribbean countries' access to disaster insurance. As a multi-country risk pool, this parametric insurance facility is able to offer disaster insurance at a significant discount relative to policies purchased on an individual basis (see Box). As shown in Table 6, the rate-on-line (the ratio of the premium to the maximum payout) is typically less than half that available from commercial insurers. Despite this deep discount, Table 6 indicates that the ECCU countries, with the exception of Grenada, are purchasing insurance coverage well below the optimum level. Indeed, using the lower rate-on-line in the simulations would yield even higher optimal insurance coverage levels at lower cost. For example, we estimate that Dominica's optimal insurance level under the CCRIF would be 25 percent of GDP with a premium cost of about 1.2 percent of GDP, compared with 20 percent and 1.8 percent, respectively, under individual commercial insurance coverage.

D. What Impact From Climate Change?

13. The frequency and intensity of hurricanes occurring in the ECCU region appear to have increased during the last decade, perhaps due to the nascent effects of climate change. As shown in Table 1, the number of natural disaster events has risen in the ten-year period to 2007 for Dominica, Grenada, and St. Vincent and the Grenadines. Data from EM-DAT suggest that half of the top ten natural disasters in the ECCU region, measured in terms of damage as a percent of GDP, occurred after 1990. Most recently, as noted in the Caribbean Natural Hazards Review 2008, the number of hurricane-strength storms in 2008 was much higher than the long-term average, with the year becoming the fourth most severe season since data has been collected.

14. The benefits of catastrophic insurance have also received greater attention given the prospects of climate change and its implications for more frequent and intense hurricane activity. Grenada is a case in point. Grenada had been traditionally regarded as located south of the Atlantic hurricane belt. Accordingly, it served as a safe haven for yachts and ships during the Caribbean hurricane season. However, the pattern of hurricane activity may have changed, possibly reflecting the effect of climate change. Following 50 hurricane-free years, Hurricane Ivan struck Grenada in 2004, causing damages estimated in excess of 200 percent of GDP, among the highest estimated loss ever from a natural disaster (see IMF 2006). The following year, another hurricane struck Grenada, this time causing damage on the order of 12 percent of GDP. In our simulation exercise in Table 5, and in contrast to the historical data in Table 3, we examine the possible impact of climate change for two cases:

  • Assuming that the costs of hurricanes for Grenada were to increase and approach the average level of its three northern neighbor counties—Antigua and Barbuda, Dominica, and St. Kitts and Nevis—then the projected maximum debt ratio in 2020 would widen to 188 percent of GDP from 144 percent, requiring a tripling of optimum insurance coverage to 15 percent of GDP per annum.

  • Assuming that climate change further heightened the vulnerability of Dominica and raised the cost of hurricanes by 33 percent, the projected maximum debt ratio in 2020 would increase to 181 percent of GDP from 150 percent. A doubling of the insurance coverage limit to 40 percent of GDP per year would help to bring down the upper bound to 132 percent of GDP.

E. Conclusions

15. The analysis illustrates that catastrophic risk insurance not only helps to reduce liquidity gaps in the aftermath of a natural disaster but also reduces resource gaps over the long run. In the case of the highly-indebted ECCU countries, insurance against natural disasters, which represents a transfer or pooling of fiscal risks, could contribute significantly to a lowering of public debt levels, and thus improve debt sustainability. Not surprisingly, our analysis also shows that countries that are most vulnerable to the occurrence of natural disasters benefit the most from insurance coverage. Moreover, this analysis attempts to determine an optimal level of insurance to improve the overall efficiency of resource allocation. This optimal level would be higher to the extent that climate change is expected to increase the frequency and intensity of hurricanes in the region.

16. While the CCRIF was designed to meet a government's immediate liquidity needs, our analysis suggests that this multi-country risk pool could play a larger role in contributing to long-term debt sustainability. Indeed, owing to the deeply discounted premia, the optimal levels of insurance would be even higher, at lower costs, than those calculated based on estimated individual insurance rates.

17. A decision to purchase disaster insurance implies balancing the amount of risk the country could assume against the fiscal space available. So far, donor assistance to pay for insurance premia has been key to enabling some of the countries to participate in the CCRIF. However, looking ahead, countries may need to fully finance their participation from their own resources. Given the ECCU countries' limited fiscal space, some may choose to forgo participation. Our analysis indicates that these countries are currently purchasing sub-optimal levels of insurance, all the more so given the conservative assumptions regarding premium costs and historical hurricane data (i.e., no impact from climate change).

18. The CCRIF is being adapted to be even more attractive to participants, thereby allowing for a potentially larger role in insuring against fiscal risk associated with catastrophes. Thus, in early 2008, the CCRIF lowered its premia and broadened its hurricane coverage. Premia were reduced by 10 percent, while maximum payouts were doubled. The insurance was also expanded to cover catastrophes expected to occur once every 15 years rather than 1-in-20-year events. Initiatives are also underway to improve the risk modeling framework and to consider adding coverage for floods/excess rainfall and agricultural damage. These changes would further raise the optimal level of insurance that should be purchased to maximize the benefits with respect to debt sustainability in the ECCU. Thus, despite the high debt levels in the ECCU, this analysis highlights the benefits to be gained by finding the fiscal space to fully participate in the CCRIF.

The Caribbean Catastrophe Risk Insurance Facility (CCRIF)

The CCRIF is a regional insurance fund that allows Caribbean governments to purchase insurance coverage to finance immediate post-disaster recovery needs.1/ The CCRIF, the first multi-country risk pool in the world, was established in May 2007, the result of a collaboration between the region's governments and key donor partners, based on a request by the CARICOM Heads of Government to the World Bank for assistance in improving access to catastrophe insurance after the severe devastation caused by hurricanes in the Caribbean in 2004. As of April 2009, the total value of the insurance facility stood at US$130 million.

Made possible by the use of parametric insurance instruments, the CCRIF provides quick claims settlement to a participating government affected by an earthquake or hurricane. Payouts are contingent on pre-established trigger events measured in terms of wind speed or ground acceleration and proportional to the estimated loss derived from a hazard impact model.

Functioning as a pooled reserve controlled by participating governments, the CCRIF retains risk from participating governments through its own reserves, and transfers risks that exceed its own capacity to reinsurance markets. The leveraging of its own reserve pool to purchase additional risk financing capacity directly in the reinsurance market allows the CCRIF to secure sufficient financial capacity to finance major losses. This structure also provides participating governments with insurance coverage at about half the price they would face if they approached the reinsurance industry independently. Other nonmembers of the CCRIF which have donated to the facility’s reserve pool are Bermuda, Canada, France, Ireland, the United Kingdom, the Caribbean Development Bank, the European Union, and the World Bank.

Insurance coverage under the CCRIF is typically capped at 20 percent of total estimated losses, a proportion which is believed to be sufficient to cover a government's immediate liquidity needs to begin emergency operations after an adverse event until other financial resources are mobilized.

The CCRIF made two payouts in its first year of operation in 2007. St. Lucia received US$0.4 million (0.04 percent of GDP), and Dominica received US$0.5 million (0.1 percent of GDP) due to damages from the magnitude 7.4 earthquake which shook the eastern Caribbean on November 29. In October 2008, the CCRIF made its first hurricane payout to Turks and Caicos in the amount of US$6.3 million due to damages from Hurricane Ike.

In light of suggestions from various stakeholders, the CCRIF is considering the possibility of insuring more frequent events and widening its coverage to include flood coverage and agricultural damage.

1/ The CCRIF currently includes 16 CARICOM members; remaining members that have not joined the CCRIF are Guyana, Montserrat, Suriname, and the British Virgin Islands.

References

  • Araujo, Juliana, 2009, “Fiscal Cycles in the Caribbean,” forthcoming IMF Working Paper.

  • Borensztein, Eduardo, Eduardo Cavallo, and Patricio Valenzuela, 2008, “Debt Sustainability Under Catastrophic Risk: The Case for Government Budget Insurance,” IMF Working Paper WP/08/44 (Washington D.C.: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Caribbean Natural Hazards Review 2008, CaribRM, Risk Managers to the Caribbean.

  • EM-DAT, Emergency Events Database (http://www.emdat.be/).

  • International Monetary Fund, 2006, “Grenada—Request for Three-Year Arrangement Under the Poverty Reduction and Growth Facility,” Country Report No. 06/277, (Washington D.C.: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Rasmussen, Tobias, 2006, “Natural Disasters and Their Macroeconomic Implications,” in Sahay, Robinson, and Cashin, eds., The Caribbean: From Vulnerability to Sustained Growth, (Washington: International Monetary Fund), pp. 122–42.

    • Search Google Scholar
    • Export Citation

  • Sahay, Ratna, 2006, “Stabilization, Debt, and Fiscal Policy in the Caribbean,” in Sahay, Robinson, and Cashin, eds., The Caribbean: From Vulnerability to Sustained Growth, (Washington D.C.: International Monetary Fund), pp. 17–57.

    • Search Google Scholar
    • Export Citation

  • World Bank, 2006, “Background Document—Initial Results of Preparation Work: Caribbean Catastrophic Risk Insurance Facility,” (Washington D.C.: World Bank).

    • Search Google Scholar
    • Export Citation

  • World Bank, 2007, “Background Document—Results of Preparation Work on the Design of a Caribbean Catastrophic Risk Insurance Facility,” (Washington D.C.: World Bank).

    • Search Google Scholar
    • Export Citation
1

Prepared by Yu Ching Wong, Anthony Lemus, and Nancy Wagner.

2

While the Caribbean countries are also exposed to earthquakes, this paper focuses on the example of hurricanes, as they represent the most frequent threat for these countries.

3

Antigua and Barbuda (ATG), Dominica (DMA), Grenada (GRD), St. Kitts and Nevis (KNA), St. Lucia (LCA), and St. Vincent and the Grenadines (VCT).

4

The return period is an estimate of the interval of time between given disaster events (such as a hurricane) of a certain intensity or size.

5

The estimates for the high probability event of 1-in-18-years are extrapolated from data points for the other return periods available from the World Bank (2006). The probable maximum loss estimated by EQECAT for Caribbean Catastrophe Risk Insurance Facility (CCRIF) countries includes estimated direct losses to government-owned assets caused by hurricanes, and estimated indirect losses due to lost tax revenue and disaster relief expenditures. Due to the lack of data such as property tax records and building code classifications, the exposure database is constructed based on field data and a series of assumptions. In this regard, the CCRIF recognizes the need to improve the exposure database in order to raise the quality of risk assessment. See also Box 1 on the CCRIF.

6

Historical data are from 1900 to 2005.

7

The stochastic shocks to each variable are assumed to follow a normal distribution with mean zero and standard deviation based on historical volatility from 1997 to 2013.

8

In the case of the CCRIF, disbursement of an insurance payout is contingent on pre-established trigger events measured in terms of wind speed (for hurricane) and ground shaking (for earthquake) thresholds. This allows the insurance payment to meet the liquidity gap immediately following the aftermath of a disaster, without the need for an on-site loss assessment, which is usually time-consuming and costly (see World Bank 2006, 2007).

9

This corresponds to an estimated commercial hurricane insurance premium and coverage ratio in World Bank (2006).

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Eastern Caribbean Currency Union: Selected Issues
Author: International Monetary Fund
Volume/Issue:
Volume 2009: Issue 176
Publisher:
International Monetary Fund
ISBN:
9781451811742
ISSN:
1934-7685
Pages:
63
DOI:
https://doi.org/10.5089/9781451811742.002
  • View in gallery

    ECCU: Estimated Cost of Hurricanes Probable Maximum Loss

    (in million U.S. dollars) by Return Period (years)

  • View in gallery

    Debt Sustainability Analysis: Central Government Debt to GDP

    (Confidence intervals at 98%)

  • View in gallery

    Debt Sustainability Analysis: Central Government Debt to GDP

    (Confidence intervals at 98%)

  • View in gallery

    Debt Sustainability Analysis: Central Government Debt to GDP– Impact of Climate Change on Dominica and Grenada 1/

    (Confidence intervals at 98%)

ECCU: Estimated Cost of Hurricanes Probable Maximum Loss

(in million U.S. dollars) by Return Period (years)

Debt Sustainability Analysis: Central Government Debt to GDP

(Confidence intervals at 98%)

Debt Sustainability Analysis: Central Government Debt to GDP

(Confidence intervals at 98%)

Debt Sustainability Analysis: Central Government Debt to GDP– Impact of Climate Change on Dominica and Grenada 1/

(Confidence intervals at 98%)