Subgame-perfect free trade networks in a four-country model

Goyal and Joshi (2006, Int Econ Review) apply the notion of "pairwise stable networks" introduced by Jackson and Wolinsky (1996, J Econ Theory) to a model of free trade network formation, and show that (i) every pairwise stable network is either complete or almost complete (with all countries except one forming direct links), and (ii) the complete network maximizes global welfare. In this note, we use essentially the same model as their model with four countries, and investigate which network is more likely to be realized than others by considering a dynamic process introduced by Jackson and Watts (2002, J Econ Theory).