This dissertation is divided into three chapters, each of which explores a particular aspect of the effect of socio-political institutions on cooperation and conflict between diverse individuals. The first chapter shows how the relationship between diversity and cooperation changes with social norms. The second chapter investigates the relationship between cultural conflict and government regulation. The third chapter describes conditions on the distribution of preferences of legislators that allow legislative coalitions to induce a legislative gridlock and block reform. In the first chapter, the notion of a 'norm of compromise' is introduced and its importance for cooperation among agents with diverse preferences over a public policy is demonstrated. Agents choose to cooperate when they join a coalition and agree to support a commonly proposed policy that could be far away from their preferred policy. A norm of compromise is an exogenous protocol used by a coalition to arrive at this commonly proposed or 'compromise' policy. I consider a parameterized class of norms in which the compromise policy's relative sensitivity to moderates and extremists in the coalition can be dialed. I study the effect of these norms on the stability of the grand coalition (or full cooperation) in a model where an agent faces a trade-off between compromise if she joins a coalition and increased risk if she does not. I find that polarization does not always reduce cooperation: it destabilizes the grand coalition under norms with low relative sensitivity to moderates but stabilizes it under norms with high relative sensitivity to moderates. This can lead to a situation where norms enabling cooperation in a polarized society do not enable cooperation in a homogeneous one. I also find the counterintuitive result that under some norms extremists are less willing to cooperate when moderates' preferences get closer to these extremists. This work sheds light on the emergence of cooperation within social movements like Black Lives Matter and the Arab Spring, where political actors compromise in the relative absence of formal institutional structures. The second chapter explores how social conflict generated through cultural diversity affects public policy. In the model, social conflict arises when culturally differentiated groups impose negative cultural consumption externalities on each other. These externalities can be mitigated by a government that uses taxation to reduce disposable income, thereby transforming cultural consumption into public good consumption. We link the size of equilibrium taxation to characteristics of the underlying distribution of cultural groups as well as to the local government's social welfare function. We test the predictions from our theoretical framework using U.S. city and county data from 1990. Controlling for a variety of socioeconomic and demographic indicators, we verify a key prediction: local taxes per capita are increasing in diversity as measured by ethnic fractionalization. We further document that other characteristics of the group size distribution have effects on local taxes per capita that are in line with our theoretical analysis. Finally, the third chapter looks at possible conditions for legislative gridlock to occur within a majoritarian assembly with endogenous coalitions. Legislators possess Euclidean preferences over multiple policy dimensions and are ordered according to a multidimensional notion of the left vs. right divide. Gridlock occurs if no coalition can propose a reform that gains stable support. As a result, the status quo is maintained. We derive both necessary and sufficient conditions for the existence of a stable gridlock. These conditions depend on two measures characterizing a legislator's ideology, which we call conservatism and extremism. Conservatism characterizes the attractiveness of the status quo relative to the median legislator's preferred policy. Extremism measures the distance of an agent's preference from the preference of the median legislator along the left-right divide. We show that gridlock can be a stable outcome only if a majority of legislators are sufficiently conservative and extremist. Moreover, gridlock occurs if a coalition that includes relatively conservative legislators located on both sides of the political spectrum tactically proposes a more conservative reform than the one that is preferred by the median legislator.