Roads for Ruling, Roads for Rebelling

• In Event: Historical Legacies of State Capacity, Conflict, and Insurgency

Abstract

State capacity figures as perhaps the most important institutional explanation of civil war. A common argument holds that state reach decays with distance from a country’s center, making conflict most likely in peripheral areas where prospective rebels are shielded from the projection state power. This paper argues that this explanation is incomplete, in that relying on a purely center-periphery dichotomy ignores the role of societal cleavages in amplifying effect of incomplete state rule. Instead, we propose that understanding the relationship between state capacity and conflict risk requires a concept of accessibility that does not just rely solely on geographical distance, but explicitly incorporates the structure of transport networks – particularly roads – among those societal groups at risk of rebelling against the state.
Road construction is an indispensable state-making tool, fostering economic, cultural, and political integration of subject populations, and facilitating the centralization of political power by giving military advantage to organized armies over local competitors. However, whether roads serve state-making goals depends on the political motives of those who build them. Importantly, in many ex-colonial countries of Africa and Asia, the backbone of today’s transport infrastructure was built by European colonial administrations. As a consequence, the emergence of road networks in these countries followed a different logic than those in other world regions. Rather than building road networks that would foster the emergence of a strong central state and eliminate societal cleavages over time – as was the case in Europe – colonial rulers were interested primarily in infrastructure that served their extractive economic interests and ensured their dominance over local subjects with minimal military overhead. As a consequence, many ex-colonial countries feature road-networks with stark regional differences in road densities and network structure, depending on distribution of exploitable resources and colonial rulers’ governance of ethnic multi-ethnic territories.
We argue that this regional variation in transport infrastructure has had long-lasting implications on cohesion of post-colonial states, moderating the degree to which these countries suffer from violent, ethno-nationalist conflict. Specifically, we argue that road-networks affect risk of a particular ethnic group rebelling against the state along two dimensions:
The first dimension is a group’s external accessibility, measuring how well a group is connected to its host country’s political center. High external accessibility has an integrating effect, increasing the frequency of interactions between group members and the state, and facilitating direct state governance in the group’s territory, thus decreasing the likelihood of ethno-nationalist conflict.
The second dimension is groups’ internal accessibility, measuring how well group members are connected among each other. Internal accessibility is conceptually orthogonal to external accessibility, but moderates its effect on likelihood of conflict: Groups with low external accessibility will be more likely to rebel the better they are connected internally, as high internal accessibility facilitates group mobilization. Groups that are tightly integrated with state center, on the other hand, will become even less likely to rebel with high internal accessibility, as efficient group-internal transport network catalyzes the diffusion of state institutions and facilitates establishment of direct control of local population.
Naturally, testing these claims systematically requires consistent, cross-national, historical road-network for the post-colonial world. Given that such data are currently unavailable, we compile a new dataset of historical road-infrastructure for the post-WWII period covering all of Africa and large parts of Asia. We do so by leveraging recent advances in machine learning to digitize two sources of archival data: (1) A collection of map sheets by the Michelin tire company covering the entire African continent in regular intervals since the 1960s, and (2) a set of declassified topological maps by the US Army Map Service covering much of Africa and Asia between 1950 and 1970.
Combining these data with geo-coded data on ethnic settlement patterns and population densities, we are able to operationalize our concepts of internal and external accessibility by using network-theoretic tools to calculate shortest-path based metrics of connectivity. Preliminary results show that accessibility measures derived from such road-network data, rather than purely geographic proxies, are indeed much better able to explain the outbreak ethno-nationalist conflict in post-colonial countries.

Authors

• Carl Müller-Crepon, ETH Zurich

• Philipp Hunziker, Northeastern University

• Lars-Erik Cederman, ETH Zurich