An Assessment of the Impact of the CARIFORUM EU-EPA on CARIFORUM Exports to the EU (2000 to 2017)

ABSTRACT

The economic literature is limited on the potential impact of the CARIFORUM-EU Economic Partnership Agreement (EPA) on exports of the region to the EU. This article applies a gravity model to assess the effect of the EPA on CARIFORUM exports to the EU for the period of 2000 to 2017. Our results show that under the EPA, the region has not been able to benefit significantly in terms of sustained merchandise exports to the EU. This observation is also made in the context of the declining complementarity between CARIFORUM economies and the EU in the post-EPA period.

KEYWORDS:

• Exports
• economic partnership agreement
• CARIFORUM
• EU

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ The EPA represents a legacy relationship which was premised on decades of the EU granting preferential access to its markets to all African, Caribbean, and Pacific (ACP) economies. There is no doubt that these preferences, though no longer available, have impacted the structure and composition of the region’s exports to the EU; this article specifically utilizes a gravity model to evaluate whether these exports were impacted in any significant way by having the EPA in place. The level of exports used in the study for the purpose of modeling is, therefore, assumed to include the impact of previous preferences on exports.

² The economic and development relations between the EU and its former colonies throughout the ACP group began in 1963 with the Yaonde Agreement and continued through to 2000 with the various Lomé Conventions and culminated with the Cotonou Agreement. The Cotonou Agreement was the first among the various agreements implemented to include the condition of reciprocal trade conditions. In the previous agreements, ACP countries were not required to remove barriers to trade such as tariffs for the benefit of EU trading partners. The Cotonou Agreement also provided the framework for the negotiation of separate agreements between the EU and various geographical blocs of the ACP countries. The CARIFORUM-EU EPA represents the Caribbean-specific agreement.

³ Mahabir (2011) was among the first to conclude that the impact of the EPA on CARICOM exports to the EU was not significant.

⁴ The authors acknowledge that the aim of the EPA was to maintain benefits alongside the conversion of unilateral preferences into a reciprocal agreement.

⁵ Other theoretical contributors to the gravity model include Bergstrand (1985, 1989) and Krugman (1980) under a monopolistic framework, Frankel, Stein, and Wei (1995) under an imperfect competition framework, Deardorff (1998) from a conventional factor-proportions description of trade, Eaton and Kortum (2002) from a Ricardian type of model, and Helpman, Melitz, and Rubinstein (2008) and Chaney (2008) from the perspective of differentiated goods with firm level heterogeneity.

⁶ To generate this graph, mirror data was used. One caveat presented by several researchers is that the region’s productive capacity is limited and so too is the scope for economies of scale, and as such, increasing the level of exports would be difficult.

⁷ This is total merchandise exports.

⁸ Note that in the period of 2008 to 2017, the number of sectors exported to the EU as a proportion of total export sectors declined for Antigua and Barbuda, Dominica, Grenada, Guyana, Jamaica, St. Kitts and Nevis, and Suriname.

⁹ The HHI is a measure of the dispersion of trade value across an exporter’s partners. A country with trade (export or import) that is concentrated in very few markets will have an index value close to 1. Similarly, a country with a perfectly diversified trade portfolio will have an index close to zero.

¹⁰ In particular, the export basket for Antigua and Barbuda, the Bahamas, Grenada, and Suriname became more concentrated between 2008 and 2017. Note that except for Suriname, CARIFORUM economies became more diversified globally. Additionally, the region’s trade with the EU is not intense.

¹¹ An aggregative trade complementarity score was estimated by reviewing the trade trends of the CARIFORUM as a single bloc. The complementarity index, as presented by Drysdale and Garnaut (1982), was used to undertake the calculation of the index. The complementarity index is calculated as $C_{ij}=\frac{{\bar{X}}_{ij}}{X_{iw}}/\frac{X_{wj}}{X_{ww}}$ where $X_{ij}$ – country i exports to country j; $X_{iw}$ – world exports of i to country j; $X_{wj}$ – world exports to country j; and $X_{ww}$ – total world exports. Note that complementarity declined between the EU and each of: Antigua and Barbuda, Dominica, Grenada, St. Lucia, Suriname, and the Dominican Republic in the period of 2008 to 2017.

¹² FC is included as a dummy with a variable of 1 for 2008, and 0 otherwise. The economic literature suggests that due to increased uncertainty, volatility, and loss of income, a financial crisis will likely result in a negative impact on an economy’s decision to engage in trade (Broll and Jauer 2014). The coefficient of this variable is found to be statistically insignificant in the outcome equation. In this context, the FC is assumed to impact a country’s decision to engage in trade.

¹³ Helpman, Melitz, and Rubinstein (2008) suggest that relative cost, the number of regulatory documents, and the number of days for trade can be used to evaluate entry conditions. Using data from the trading across borders sub-index of the overall doing business rankings for exporters (CARIFORUM) and the importer (EU), we construct a pairwise index that takes the values between 0 (very low cost of market entry) and 3 (very high cost of entry).

¹⁴ For the OLS and HSS models, note that all continuous variables in the model are included in log form (i.e., the reported coefficients are elasticities) while for the PPML only, the independent variables are logged.

¹⁵ The findings of the selection equation are not presented for brevity but are summarized at the end of Table 2.

¹⁶ The University of the West Indies (UWI 2013) also showed that the coefficient for the EPA variable was −0.67 (in 2012).

¹⁷ Mahabir (2011) also returned a negative coefficient for the POLS and fixed effects permutations of the specified model.