Ethnic conflict, rumors and an informed agent

Abstract

Rumours often precipitate ethnic conflicts. There may exist an agent (player b) who knows if the rumour is true or false. We explore the role that such agents play in negating the effects of a bad rumour. Two cases are analysed - b is non-strategic i.e. he reveals his information truthfully and b is strategic and has a commonly known bias towards one ethnicity1 . There are two states of the world. One in which conflict can be avoided (good state) and one in which it is inevitable (bad state). Rumour is a public signal which is correlated with the bad state of the world and therefore precipitates a conflict situation. Before deciding to fight or not, there is a meeting stage in which player b (if he exists), can send signals to k percent of the population. We find that when b is non-strategic, conflict caused by a false rumour is unlikely to happen when b can meet a large fraction of the population (as in rural areas). This is due to two effects - not only do the players who meet b know that the rumour is false, they also estimate (from the commonly known meeting process) that a large part of the population must also know. This allows them to coordinate, not fight and enjoy the high peace time payoff as opposed to the lower conflict payoff. This could be a possible explanation for the commonly observed phenomenon that ethnic conflicts in India are largely an urban phenomenon2. When b is allowed to be strategic and it is known that he has a bias towards one ethnicity, we show that, under parametric restrictions, there are only two equilibrium outcomes possible. In one outcome, everyone fights and conflict occurs with probability one. In the other equilibrium outcome, there may not be conflict when b believes that peace is possible and when b believes that conflict is inevitable, he reveals information in a way which gives his own ethnicity a higher probability of winning.