Chapter I: Cohesive Anarchy --- In a conflict, a small force may overcome a larger one if the latter fails to coordinate. To understand how incentives drive such failures, I study Tullock contests between a cohesive faction and a group embedded in a network. The group’s strength equals the sum of its members' efforts, where links measure their pairwise complementarity or interference. I characterize equilibria for general networks, and find only few members are likely to contribute. Furthermore, the prize and a network measure of interconnectedness jointly improve the internalization of spillovers by the group, so its performance varies with the stakes -- escalation induces cohesion.

Chapter II: Spectral Oligopolies --- We study how demand interactions incentivize multiproduct oligopolists to reduce their variable production costs. Our starting point is an equivalence between multimarket competition over dependent products and single-market competition over independent bundles from which we derive three insights. First, heterogeneous innovativeness begets a core-fringe separation. While strong innovators dominate clusters of complementary markets through demand-side economies of scope, others retreat towards niches whose isolation attenuates the impact of investments. Second, the translation from innovative advantage into profits is strongest in `contractor' graphs, the antipodes of expanders. These graphs comprise clusters which are densely linked within but scarcely without, yet synergize well both by themselves as well as collectively. Consequently, the parts and the whole make for attractive and relatively independent investment targets. Third, as demand interactions scale up, both the market concentration and consumers' share of the surplus rise under broad conditions, so market-share based indices of concentration tend to suggest losses for consumers when the opposite is the case. We construct a generalized Herfindahl index which overcomes this limitation.

Chapter III: Dominant Firms --- Many consumer industries evolve into partial oligopolies where firms with and without market power coexist. I develop a framework of dynamic competition which explains this pattern. Starting from a market with a continuum of firms, companies stochastically adjust their product offerings through the accumulation of thin-tailed innovations. In conjunction with discrete-choice founded demand, disruption becomes possible: if the spread of tastes relative to innovations' volatility falls below a threshold, occasionally a firm separates from the continuum with an outstanding product. As a result, this dominant firm accrues a positive market share and profit margin until a future innovator supplants it. During these cycles of disruptive turnover, the fringe provides a `seedbed': since incumbents emerge as frontrunners of the ongoing race of innovations, their products' qualities increase in the number of candidate firms from which they are drawn. Through this mechanism, the fringe’s measure is a first-order generator of surplus over time even if its output is not.