Revealed Multilateral Trade Resistance in Gravity Models - A Network Approach

The position of countries in the world trade network (WTN) is fundamental in understanding and explaining trade flows between countries. Recent theoretical models introduce these network structures into the traditional gravity model, most notably Anderson and van Wincoop (2003) by introducing the concept of multilateral trade resistance (MTR). Despite these theoretical insights, little has been done to empirically validate these network effects. The concept of MTR remains a black box with many unobservable factors. The key contributions of this paper are the following. First, we capture the network effect of the WTN by the notion of revealed multilateral trade resistance (RMTR), which consists of calculating first-order and second-order network effects of the WTN. Secondly, we show that this network structure is important to explain bilateral trade patterns, by presenting an augmented gravity model, which includes the RMTR. In particular, we characterize the RMTR by two network indicators, degrees and clustering, and show that these indicators have strong and significant, but opposing effects on bilateral trade values. A higher degree raises average bilateral trade values, while a higher clustering coefficient has a negative impact on bilateral trade. Finally, we decompose the effect of the RMTR by studying its effect on the intensive and extensive margins of trade. We find that the effect of degrees is positive and fairly symmetric on both the intensive and the extensive margin of trade. The effect of clustering is, however, positive on the intensive margin, but negative on the extensive margin, indicating competition effects. Hence it appears that countries decrease their variety in exported goods due to competition in the global network. However, they trade their varieties more intensely