Introduction

International trade flows have increased impressively in the last few decades. Data from the United Nations, which have been edited by Feenstra et al. (2005), suggest that the aggregate nominal value of reported international trade increased from about US$130 billion in 1962 to more than US$6.5 trillion in 2000. This corresponds to an annual growth rate close to 11 per cent. With an estimated world population of about 6 billion in 2000, this implies that international trade per capita was over US$1,000 or about 15 per cent of the average Gross Domestic Product (GDP) per capita.

The growing importance of international trade has led to an increased need for sound analyses of its determinants. The gravity model has been the workhorse model to explain international trade flows for nearly half a century now. The main idea behind this model is that the magnitude of bilateral trade flows can be explained by the economic size of the two trading countries and the distance between them (Deardorff 1998). The model has sound theoretical foundations, yields almost invariantly plausible parameter estimates, and has a strong explanatory power.

Although the basic framework of the gravity model is unaltered throughout the years, new insights have contributed to its increasing popularity by improving its theoretical underpinnings (see, e.g. Feenstra 2004) and addressing econometric issues concerning the correct specification of the model (see, e.g., Anderson and van Wincoop 2004).