We examine the ongoing transition from centrally planned to market agriculture in rural China. In particular, we examine the devolution of land rights from village governments to villagers and the corresponding evolution of tenure security in agricultural land. We find econometric support for the statistical and economic importance of four explanations for local government behavior. Three of these explanations indicate a link between the incentives and constraints faced by village leaders and property rights in agricultural land, and hence suggest policy levers to encourage more secure property rights.

Rising extremism and polarization threaten democratic institutions worldwide. As opposing factions become more extreme in their opinions, polarization increases— the chasm widens the chasm between fellow citizens, and common ground erodes, washed away down a river of vitriol, bitterness, and hate. What causes increased extremism and polarization? Due to the highly complex nature of human soci- eties, this problem of explaining polarization must be broken down into many sub- problems, which themselves require complex systems thinking to address. Simpli- fied models of social systems and rigorous analysis of empirical data, are necessary to build a thorough, coherent understanding of social behavior.In this dissertation I present my findings from studying three sub-problems in explaining why and how extremism and polarization emerge. First, I focus narrowly on a communication strategy shown in behavioral studies to increase extremism, metaphorical violence, such as “Biden hit Trump over his tax returns in yesterday’s debate.” While we know the effects of violence metaphors, we do not understand their distribution in the wild, or what causes their usage to increase and decrease. I found that metaphorical violence use increased around the time of presidential debates and elections in the United States, and was correlated with presidentical candidates’ tweets. Second, I show that rising extremism in isolated social gropus may be simply explained by the fact that extremists are more stubborn than centrists—however behavioral studies on the subject may contain ubiquitous false detections of rising extremism, which I demonstrate in Study 3. Finally in Study 4, I developed and analyzed an empirically motivated, network theoretic, agent-based model of social influence at the societal level to understand how well we can predict polarization based on the effects of initial conditions, network structure, communication noise, and random chance on predictions of polarization. Taken together these studies advance our understanding of communicative, cognitive, and social factors in the emergence of extremism and polarization.