Fitting the gravity model when zero trade flows are frequent : a comparison of estimation techniques using Africa's trade data

The issues of zero trade observations and the validity of the log linear transformation of the gravity equation have generated a number of debates in the literature with differing claims about the most suitable estimation technique. To produce unbiased and consistent estimates for policy making, we undertake a careful comparison of a number of widely used estimators to investigate if EU fish standards are protectionist following reoccurring rejection of African fish products at the EU border. Analysis was based on a dataset of Africa's fish exports to the European Union between 2007 and 2012, which contains about 63% zero trade observations. Our results from the robustness checks are in favour of only the Multinomial Poisson Maximum Likelihood (MPML) technique as the most consistent estimator in relation to the impacts of standards and other explanatory variables. In addition, we find EU standards are indeed non-protectionist in spite of the high level of African fish exports rejected since 2008 at the EU border. Thus, a deeper trade agreement between these trading partners involving a significant transfer in science and technology to the Africa could help improve their compliance rate to EU standards and ensure increased export penetration.