We develop a model to study coalitions that extract the resources of outsiders. The players in our model are endowed with power and resources. The ruling coalition plunders outsiders, distributes the plundered resources among its members, and guarantees that insiders’ resources remain safe. Our analysis focuses on the resilience of the equilibrium ruling coalition to exogenous shocks affecting the power and resources of both insiders and outsiders, as well as the intensity of plundering. We show that hierarchy—an internal ordering in which members are equal within ranks and distinct ranks are weakly ordered in both power and resources, with at least one strict inequality—weakly increases external resilience to shocks to outsiders’ power and resources. The gain is strict when a relevant insider sub-coalition is locally decisive; a sufficiently low uniform bound on the marginal return to power provides a sufficient condition for local decisiveness. By contrast, under our uniform large-slope condition, the external boundary depends only on the ruling coalition’s aggregate power and resources, so internal redistribution has no effect. We interpret these slope conditions as capturing weak and strong property-right protection. Our final result identifies a condition for a trade-off between internal and external resilience.

We study how the timing of intervention affects optimal targeting in network games with strategic complementarities and incomplete information. Individuals differ in their number of connections and choose effort before knowing who their neighbors will be. A social planner allocates a fixed enforcement budget across degree classes to minimize total surviving activity. When enforcement and effort are chosen simultaneously, the unique equilibrium targets the most connected individuals and equalizes effort across targeted classes, without fully eliminating any class. When the planner can commit to enforcement before individuals choose effort, however, it anticipates how intervention reduces effort not only among those directly targeted but also throughout the network. We show that these indirect effects, captured by a measure of network amplification, can make it optimal to concentrate enforcement more heavily on highly connected individuals. In particular, when the importance of a degree class in individuals’ beliefs about their neighbors, relative to its population size, weakly increases with degree, network amplification is stronger for more connected individuals. Under this condition, whenever simultaneous intervention targets at least two degree classes, reallocating some enforcement toward the more connected class strictly reduces total surviving activity. Finally, we identify an exact enforcement-budget threshold above which the ability to commit strictly improves outcomes relative to simultaneous intervention.

We construct a general axiomatic approach to measuring spatial concentration around a center or capital point of interest. Building on expected utility theory, we propose a basic axiom of independence (subgroup consistency), and show that it implies an expected influence representation of the concentration order. We prove that two additional axioms (monotonicity and rank invariance) imply that the associated influence function is a decreasing isoelastic function of the distance to the capital. We apply our index to measure the concentration of population around capital cities, show its advantages over alternative measures, and explore its correlations with many variables of interest.

Many access reforms are designed to broaden participation by lowering the cost of becoming an effective competitor. This paper shows that such reforms can have the opposite effect when access is contestable. I study a contestable-access model with ratio-form payoffs in which potential claimants differ in the cost of producing effective contest intensity. A uniform reform helps marginal claimants by lowering their cost of becoming effective competitors, but it also strengthens low-cost incumbents. When incumbents respond strongly, aggregate contest intensity rises and the endogenous participation cutoff moves against marginal claimants. Participation then falls even though every claimant faces a lower cost. I characterize this reversal using a lower-tail concentration index that measures whether effective intensity is concentrated among the lowest-cost active claimants or spread toward the participation margin. A uniform reform expands participation when intensity is spread toward the margin and contracts participation when intensity is concentrated among low-cost incumbents. I then derive primitive curvature conditions on the contest cost function that determine which of these two cases obtains. The paper also compares uniform reform with targeted entry expansion. A policy that targets claimants near the participation cutoff has a first-order effect on entry, while its equilibrium cutoff feedback is second order. Small budgets spent at the boundary therefore generate larger access gains than equal-budget uniform reforms. The results explain why broad administrative simplification may fail to expand access in contestable allocation systems, and why targeted support can be more effective when the objective is participation.