Gravity Model of German Trade: Poisson Regression Approach

This paper utilizes the gravity model to analyse German exports. We have used a panel dataset which includes German exports to 176 countries and 20 control variables, including institutional factors during the period of 1995-2011. We estimate the model by applying a Poisson regression with robust variances and estimate three models (a pooled Poisson regression as well as a Exed and random panel data Poisson regression). The results reveal that the core economic variables (partner GDP, distance) are the most relevant factors behind German exports. The distance coefficient is slightly below the average results from other studies and thus in comparison German exporters are less harmed by geographical distances than exporters in other countries. The real exchange rate has also been revealed as relevant: depreciation stimulates exports. Other economic significant variables were population (negative effect), recession (negative effect) and domestic GDP (positive effect). The results of institutional variables are not conclusive regarding their effects on exports.