Influences of Sociocultural Networks on the International Trade Network: Evidence from the Exponential Random Graph Model


Min Zhou, University of Victoria

The international trade network is one of the most prominent manifestations of economic globalization. It reflects complex interconnection and interdependence among national economies. Thanks to its importance it has become one frontier of the scholarship on the intersection of social network analysis (SNA) and economic globalization. Despite many advances made in the application of SNA to the international trade network, empirical research on sociocultural influences on the international trade network has been scarce. The theoretical foundation of sociocultural influences on economic activities arguably originates from Coase’s (1937) transaction costs theory, but it is economic sociologists who have made systematic contributions. Granovetter (1985) introduced the idea of “embeddedness” that sees economic relations as embedded in real social networks rather than abstract idealized markets. Economic sociologists have illustrated the importance of sociocultural influences on various economic activities (Smelser and Swedberg 2005). Such theories as transaction costs and embeddedness provide a theoretical foundation for a social dimension underlying the international trade network. Nevertheless, empirical research has been scarce in actually revealing what and how sociocultural factors shape the international trade network. There are some notable exceptions (Frankel 1997, 1998; Zhou 2010, 2011). These exceptions confirm various sociocultural influences in shaping international bilateral trade, not the international trade network, though. According to these studies, sociocultural influences on bilateral trade have been on the rise over time. Countries are increasingly attracted to socioculturally similar countries when conducting international trade. As a result, international trade displays a tendency towards regionalization along sociocultural lines. Nevertheless, when this scarce existing empirical literature examines sociocultural influences on international trade, it predominantly employs the gravity model borrowed from international economics, instead of SNA. There are major limitations of this approach. The gravity model explains bilateral trade flows but cannot account for the overall structure and formation of the international trade network. In other words, the gravity model remains at the dyadic level, rather than the network level. It treats bilateral trade as independent from each other and thus ignores interdependence of bilateral trade in the international trade network. Bilateral trade is not simply a business of the two countries involved, but is also under systemic influences from the overall international trade network. The SNA is more effective in taking into account systemic influences than the gravity model. Consequently, there is a notable gap in the scholarship on the international trade network. This study is a first attempt to fill this gap and foreground sociocultural factors when applying SNA to international trade. It employs SNA tools to examine various sociocultural influences on the structure and formation of the international trade network. Specifically, we use the Exponential Random Graph Model (ERGM) to investigate influences from five major international sociocultural networks (the common language network, the common religion network, the historical colonial network, the regional trade agreement (RTA) network, and the common currency network) on the international trade network. We distinguish two types of such influences, the embeddedness effect and the positional effect. They represent two distinct mechanisms through which international sociocultural networks affect international trade. The embeddedness effect focuses on the direct influence of sociocultural connections, whereas the positional effect captures the indirect influence from a country’s position (centrality) in international sociocultural networks. The ERGM modeling of international trade data in 2010 generates interesting findings. First, connections in the common language, common religion, and RTA networks all significantly promote the formation of trade relations, whereas connections in the historical colonial and common currency networks show no effect. Second, positions in different sociocultural networks display differing effects on the international trade network. While a more central position in the RTA network promotes a country’s trade relations with others overall, more central positions in the common language, common religion, and especially common currency networks may actually depress the formation of trade relations with other countries in general. A more central position in the historical colonial network shows no significant effect. We further discuss the explanations and implications of these findings.

Non-presenting author: Gang Wu, Southwestern University of Finance and Economics

This paper will be presented at the following session: