Abstract
The gravity model that serves as an important framework to explore the relation between exchange rate volatility and international trade suffers from two weaknesses: selection bias caused by dropping of observations with zero trade flows and the inability to predict asymmetric bilateral trade flows. The latter includes situations of bilateral trade in one direction but not in the opposite direction. While some recent literature has addressed the selection bias, there are no studies that address both problems in the context of the effect of exchange rate volatility on international trade. The article contributes to the literature by applying a recent model of firm selection to control for both biases. We found that the effect of exchange rate volatility on trade, that appears to be weak under the standard gravity model or in those models that only correct for sample selection bias, emerges as a strong negative effect as both biases get controlled.
Notes
1 Henceforth HMR.
2 See Helpman, et al., (Citation2008) for detailed derivation of the equations in this section.
5 The Religion variable is an index capturing the similarity of religious composition between two countries. It is calculated as (% Catholics in country × number of Catholics in country) + (% Muslims in country × number of Muslims in country) + (% Protestants in country × number of Protestants in country).
9 We define unbalanced exports as {Absolute(Exports ij – Exports ji )}/(Exports ij + Exportsji).
10 All estimations presented in include exporter, importer and time-fixed effects.
11 These two terms are the sources of nonlinearity in the system.