Macroeconomic fluctuations in the Eastern Caribbean: The role of climatic and external shocks

Abstract

This article develops country-specific vector autoregressive (VAR) models with block exogeneity restrictions to analyze how exogenous factors affect business cycles in the Eastern Caribbean. It finds that external shocks play a key role, explaining more than half of macroeconomic fluctuations in the region. Domestic business cycles are especially vulnerable to changes in climatic conditions, with a natural disaster leading to an immediate and significant fall in output – but the effects do not appear to be persistent. Oil price and external demand shocks also contribute significantly to domestic macroeconomic fluctuations. An increase in oil prices (external demand) is contractionary (expansionary), and the effects dissipate up to three years after the shock.

Keywords:

• Caribbean
• spillovers
• supply shocks
• VAR model

JEL Classifications:

• C32
• E32
• F41

Notes

 1. The Eastern Caribbean countries analyzed in this article are the six Fund members of the Eastern Caribbean Currency Union (ECCU): Antigua and Barbuda, Dominica, Grenada, St. Kitts and Nevis, St. Lucia, and St. Vincent and the Grenadines.

 2. See Cashin (2004, 2006) for a characterization of business cycles in the Eastern Caribbean.

 3. The block exogeneity procedure reduces the number of parameters to be estimated, limiting the erosion of degrees of freedom and improving the efficiency of the estimation.

 4. Rasmussen (2004, 2006) argues that small island states are especially prone to natural disasters, with the countries of the ECCU standing out owing to the large number of hurricanes that strike the region. He notes that a natural disaster typically occurred once every 4.5 years in each of these countries.

 5. Robustness checks reveal that the main results described in this article do not change substantially when using alternative orderings of these variables.

 6. CRED is the most comprehensive database on natural disasters that is publicly available. See Rasmussen (2004, 2006) for additional details on how natural disasters are defined, and see the Appendix for details on the timing of natural disasters in each of the ECCU countries.

 7. The literature on external shocks in small open economies has used both dummies and continuous variables to measure the incidence of natural disasters and climatic factors. For instance, Raddatz (2006) uses dummy variables to capture the incidence of natural disasters, while Buckle et al. (2002) use a continuous measure of soil moisture conditions to capture the impact of changes in climatic conditions.

 8. Standard unit root tests (augmented Dickey-Fuller) show that all the variables are stationary in first differences. In addition, co-integration tests (Johansen) suggest that the variables in the model are not co-integrated. Hence, it is adequate to estimate the model in first differences.

 9. The specification of the estimated equations follows from the block exogeneity restrictions discussed earlier. The model is estimated with two lags – the lag length being selected according to the Akaike Information Criterion (AIC).

10. Throughout this article, impulse response functions trace out the response of domestic output to one standard deviation shocks, with standard error bands representing plus/minus two standard errors. In addition, variance decomposition analyses focus on a horizon of three years, unless otherwise indicated.

11. This broad characterization of the importance of external shocks is consistent with the openness of these tourism-dependent economies (see IMF 2008, 2009).

12. Shocks to GDP growth in advanced economies can be transmitted to the ECCU countries through other channels, such as the substantial inflows of remittances resulting from the large-scale migration of skilled workers to industrial countries.

13. The correlation coefficient between GDP growth in industrial countries and that of the ECCU is 0.71.

14. Sosa (2008) finds that shocks to US growth are passed on to Mexico almost one-to-one as well. The fact that elasticities for the ECCU are higher is notable, given the large trade linkages between Mexico and the United States.

15. Specifically, the benchmark specification was modified to include foreign aid as the last variable (in the Choleski ordering) of the international economy block. The variable used is OECD's Official Development Assistance (ODA).

16. See, for instance, Easterly (2003), Easterly et al. (2004), and Rajan and Subramanian (2008).

17. Cashin and Dyczewski (2006) provide a discussion of the major channels through which small, disaster-prone economies can respond to the challenge posed by natural disasters, focusing on the options for mitigating disaster risk in the Caribbean.