Natural resources often provoke conflicts within societies (Collier, 2007). This presentation analyzes the interaction between institutional quality and natural resources. The first part presents a theoretical investigation of revolutions in resource-dependent countries. Using an experiment, the second part then provides a behavioral perspective on conflict.

In the first part, "Resource exploitation, leader behavior, and the timing of institutional change", a modified AK model of revolutions (e.g. Arab Spring) is proposed. This dynamic framework allows for the endogenous determination of the timing of institutional change. A resource-dependent society is assumed to be composed of an incumbent elite and the opposing citizens. In the initial regime, the elite benefits from absolute control over resource wealth. Under elite dictatorship with a revolutionary threat, the elite always transfers a fixed fraction of the resource to the citizens. The decision of when to conduct a revolution is decided by the citizens. A revolt results to a second regime of equal access over the resource. The citizens have to decide whether or not to instigate costly conflict, and when. Different from previous literature, this paper’s integral contribution lies on the explicit determination of the timing of revolt. This is done using a differential game setting where the elite chooses a transfer policy. Under growth-friendly elite behavior, a revolt always occurs when the citizens’ revolutionary cost (i.e. repression extent) is low. When the cost is high, the elite is able to provide transfers sufficient to hinder a revolt. Finally, this setup is modified by considering military spending. Preliminary numerical analysis show that, under certain circumstances, military spending (rather than transfers) is preferred by the elite.

In the second part, "Interplay between common-pool size and endogenous institutional choice", conflict over natural resources is explored in a controlled laboratory setting. The experiment provides an investigation of the resource curse: "countries with larger endowments of natural resources tend to experience less development." It is based on a modified rent-seeking game. Subjects were assigned to groups to play two sequences (low or high level of resource). They were given an endowment that they had to allocate between a private activity and a resource rent-seeking activity. Baseline results show that a larger fraction of groups’ aggregate endowment is invested in rent-seeking when the resource level is high. The test treatments then explore the impact of endgenously determined institutional protection of the resource. In the dictatorship treatment, only one member of the group chooses how much of the resource is protected from rent-seeking. This is compared to the extreme setting, majority voting, where all members decide. Individual behavior data from the dictatorship treatment show that, in some cases, dictators rent-seek more than non-dictators. Indeed, it was shown that majority voting produces more efficient outcomes.