Cooperation and conflict are central to economics and to human life in general. One of the most basic choices we must make in almost any situation is when to collaborate and when to compete, when to help and when to harm. This fundamental question of cooperation versus conflict recurs in a variety of domains, ranging from individuals to entire nations. Yet despite the many differences between these domains, certain common threads unite them, and one of my primary research goals has been to explore these common threads. In the first chapter of this dissertation, I consider the question of cooperation and conflict at the smallest scale, between individuals in a group. One of the simplest experimental representations of cooperation is the public goods game, which has been a staple of experimental work for decades. I devised a twist on this classic experimental design. Past public goods experiments have treated taxation as pure waste or assumed flat taxes. I study a more realistic setting in which taxes are progressive, tax revenue provides a public good that may be more or less efficient than contributions, and contributions are tax deductible. Though there is an interior dominant strategy equilibrium, mean contributions are consistently above this level, even when this is Pareto-harming. The motivation to contribute is resilient in the face of tax incentives---even when such ``cooperation'' is harmful to everyone. In the second chapter of this dissertation, I consider the question of cooperation and conflict at the largest scale, between nation-states. While national conflict is the primary subject, it is worth nothing that the very existence of functional nation-states requires cooperation on an enormous scale. Thus, war is both a failure of cooperation and also an exemplar of it. The central phenomenon I sought to better understand is the present rarity of international conflict: Wars are less frequent today than through most of history. Various explanations for this have been proposed, ranging from the spread of democracy, to globalized trade, to the invention of nuclear weapons. To investigate this question, I develop an economically microfounded model of international conflict as an indefinitely iterated game, showing that under typical parameters there exist many equilibria, some resulting in war, others in peace. I propose that international norms provide a plausible equilibrium selection mechanism for this iterated game. This offers a synthesis between the ``realist'' and ``liberal'' schools of international relations: The realists are correct that international norms are cheap talk, not written into the rules of the game; but the liberals are correct that norms can have real effects, when they select one equilibrium over another. In the third chapter of this dissertation, I and my collaborators consider the question of cooperation and conflict on a more moderate scale: tacit collusion between firms in an oligopoly. This case is interesting because it is a form of cooperation on one scale that entails harm on another scale: Consumers would most likely be better off if firms chose not to cooperate with one another. In particular, we hypothesized that tacit collusion may provide an explanation for the phenomenon of Asymmetric Price Transmission (APT), the tendency of prices to respond more rapidly to positive than to negative cost shocks. Using a laboratory experiment that isolates the effects of tacit collusion, we observe APT pricing behavior in markets with three or more sellers, but not in duopolies. Furthermore, we find that sellers accurately forecast others’ prices, but nevertheless consistently set their own prices above the profit-maximizing response, particularly in the periods immediately following negative shocks. Overall, our findings support theories in which tacit collusion plays a central role in APT. A few common insights can be gleaned from these three different lines of research: First, cooperation is a very common outcome, even when circumstances are contrived to make it especially difficult. Cooperation only occurs in small space of possible equilibria in the international conflict model, yet in the real world cooperation is the most common outcome. In the experimental studies, matters are even more extreme: as these are finitely iterated games with a unique stage-game Nash equilibrium that is uncooperative, cooperation is theoretically not sustainable as an equilibrium at all. Yet on average we saw more cooperation than conflict in both experiments, and most participants in most sessions chose cooperatively at least some of the time. Second, human beings are not perfect cooperators; the tension between conflict and cooperation is a persistent one. Obviously wars do occur in reality, even though they are rarer today than they once were. In experimental studies, cooperative behavior is very common, but so is uncooperative behavior, and most choices lie strictly between the optimal cooperative and competitive choices. Third, the conditions under which cooperation can be sustained are subtle, complex, and often difficult to explain. This was particularly apparent in the public goods experiment, where neither self-interest, warm-glow altruism, nor inequity aversion could adequately explain the observed pattern of behavior. It was also apparent in the international conflict model, where surprisingly subtle changes in economic parameters could make a peaceful equilibrium collapse into war. There remains a rich unexplored frontier of research in better understanding why and how human beings engage in cooperation and conflict.