show that increasing distance reduces trade more than proportionally. Both the GDP
variables and distance are Statisticallyhighly significant7.
In the third specification we add dummy variables that are meant to capture other
bilateral factors such as geographical adjacency, and shared political and cultural
backgrounds. Following the change in specification, the coefficient on geographical
distance falls somewhat in absolute value. This reflects the negative correlation between
distance on the one hand and especially common border and free trade agreement on the
other. The contribution of the dummies to explaining the size of trade is substantial in
economic terms, although they are not as important statistically as GDP or distance.
Especially membership of both countries in a common free trade agreement has a large
discrete effect on estimated trade. Two members of a trade block, on average, trade 146%
more than two countries that do not take part in the same trade block, all else equal8.
Alternatively, a common language implies that estimated trade rises by 27%. Our results
confirm another noticeable pattern often found in the literature. Countries that are directly
adjacent trade much more, irrespective of the geographical distance between the countries.
The estimated adjacency effect amounts to 79% extra trade, ceteris paribus. It reflects a
concentration of international trade on a sub-national, regional level, clustering
substantially along borders, partly because of the strong distance dependence of trade. This
implies that the distance between the two centres of gravity of neighbouring countries
overestimates the average distance of trade between them. The argument that the distance
measure used leads to an overestimate of the distance of trading holds true for all pairs of
countries. However, its relative impact is much larger in neighbour countries than in
countries that are far away from each other. A part of the border effect may be related to
factors such as good infrastructure and strong cultural, ethnic and other historical ties along
borders that are not reflected by other variables. Also the existence of differences in
regulation between neighbouring countries may lead to an increased intensity of
international trade along the borders, because of the opportunity to take advantage of
arbitrage possibilities.
The effect sizes of all bilateral variables are statistically significant at the 5%-level. In
the remainder of the paper we will depart from specification 3 to investigate the
contribution to the explanation of trade flows of additional country-specific and bilateral
Bilateral Trade Flows and Institutions