Visualizing the Regionalized Structure of Mobility between Countries Worldwide

Abstract

Despite the sociological relevance of human mobility between nation-states, our knowledge about its planet-scale structure is still limited. Here, geographic mapping, algorithm-based community detection, network visualization, and conventional line plots are combined to display the network structure of 2.3 billion estimated trips between countries worldwide in 2016, together with information about the (non)evolution of this structure over time. The graph reveals that transnational mobility is highly regionalized: 80 percent of all human movements between countries occur within world regions. Despite strong increases in the absolute amount of transnational mobility, this share remains extremely stable between 2011 and 2016. The community detection algorithm reveals six mobility clusters that clearly correspond to world regions: Africa, Asia and Oceania, the Americas, Eurasia, Europe, and the Middle East. This stable, regionalized structure suggests that a fully globalized “world society” is unlikely to emerge, as social ties remain parochial, even in the transnational sphere.

Human mobility that transcends national borders has many sociologically meaningful consequences, from spawning cosmopolitan attitudes (Mau, Mewes, and Zimmermann 2008) to causing new fault lines (Jeannet 2020). Indeed, Simmel’s ([1908] 1971:23) classic argument that “society exists where a number of people enter into interaction” implies that when people are transnationally active, they create “society-building” ties between nation-states. But how are these ties structured at the planetary scale today? Do they form a fully globalized pattern, as terms like global village suggest, a core-periphery structure in line with world-systems theory, or rather a regionalized structure—a “world of regions” with sparse connections between continents?

Figure 1 provides an answer. It maps 2.9 billion estimated transnational trips between 196 countries worldwide in 2016, on the basis of the open-access Global Transnational Mobility Dataset (Recchi, Deutschmann, and Vespe 2019). A community detection algorithm is used to automatically find clusters of countries that are well connected to one another through human movements, while mobility between clusters is sparse. Six clusters are retrieved (marked by different colors in Figure 1), all of which are clearly congruent with world regions: one cluster almost exclusively contains countries in the Americas (pink nodes), another is mainly European (orange). Then there are Eurasian (red), Middle Eastern (blue), Asian and Oceanian (dark green), and African (light green) clusters. Intracluster ties between continents are rare exceptions that seem driven partly by the retroactive structuring forces of colonialism in combination with a secluded geographical location. ¹ Yet overall, the most salient feature is the strongly regionalized structure of transnational human mobility.

Figure 1.

The regionalized structure of human mobility between countries worldwide. In the network, nodes represent countries (positioned according to the geocoordinates of their capital), and the ties between them are estimated trips, on the basis of the Global Transnational Mobility Dataset (Recchi et al. 2019). The tie width corresponds to the number of trips (see, e.g., the United States and Mexico or China and Hong Kong for particularly large mobility flows). Only ties with more than 5,000 estimated trips are shown. The node size corresponds to the weighted degree centrality (i.e., the number of in- and outgoing trips) of the country. The colored clusters denote groups of countries that are particularly well connected to one another through trips and have been detected by a modularity-based community detection algorithm (see the Supplemental Material for further details). Inset A shows the growth of human mobility between countries within world regions (categorized via the United Nations M.49 GeoScheme) between 2011 and 2016. Inset B shows a similar trend for interregional mobility. Inset C reveals that at 80 percent, intraregional mobility as a share of all mobility is extremely stable over the entire time period under study. This structural stability results from the fact that the growth rates of mobility within and between regions are exactly the same, namely, 4.8 percent per year on average.

How has this network evolved over time? Inset A shows that within world regions, ² the amount of transnational mobility grew at a remarkable pace between 2011 and 2016, the time frame for which the data are available. The quantity of mobility between world regions (inset B) is much lower but features precisely the same high growth rate, namely, 4.8 percent per year on average. As a consequence, the regionalized structure is extremely stable over time, as inset C reveals: in each and every year from 2011 to 2016, precisely 80 percent of all transnational trips took place within world regions. Thus, whereas the network’s size has increased by more than a quarter in just five years, its structure remains unchanged.

This remarkable structural stability—which for individual mobility types such as migration has been shown to go back until 1960 (Deutschmann 2016)—has important sociological consequences. For example, it implies that the social bonds that are built through cross-border interaction (in line with intergroup contact theory) are primarily created between countries within the same world region. As a consequence, they are unlikely to contribute to the creation of global, truly cosmopolitan solidarity, trust, and sense of community. Rather, regionalized interaction networks are prone to produce “regiopolitan” mind-sets.

Of course, occasional cross-continental ties do exist (and are indeed also visible as thin lines in the background of Figure 1). Technically, we certainly live in a “small world”: the average path length between two countries in the network was only 1.69 in 2016 and thus much shorter than the famous “six degrees of separation” of the small-world phenomenon (cf. Watts 1999). However, when we take not just the existence of ties into account but also their weight (i.e., the number of estimated trips), the picture looks different. Short paths between distant corners of the globe do exist, but most of the time, humans move along the beaten paths within world regions. By and large, mobility—and the social bonds it triggers—remains parochial and regionalized, even in the transnational sphere.

Supplemental Material

VisualizingRegionalMobility_FINAL_SupplementaryMaterial – Supplemental material for Visualizing the Regionalized Structure of Mobility between Countries Worldwide

Supplemental material, VisualizingRegionalMobility_FINAL_SupplementaryMaterial for Visualizing the Regionalized Structure of Mobility between Countries Worldwide by Emanuel Deutschmann in Socius