Essays on Information and Political Economy

Abstract

This dissertation consists of three essays on media, political and learning. More specifically, I investigate the effects of biased media and learning from that biased media on political institutions. In the first essay, titled “Optimal Dynamic Information Supply and Competition”, I provide a model of an information market where the viewers acquire signals each period at an attention cost, solving an optimal stopping problem à la Wald (1947), and the objective of the potentially biased information providers is to maximize the number of viewers who acquire signals from them across periods. I find that, in a monopoly market, the information provider sends unbiased signals that perfectly reveal the state of the world when there is a single period but provides biased signals when there are multiple periods. This is because biased signals elongate the learning process of some viewers, potentially increasing the information provider payoff. I also find that incentives due to competition, modeled as another information provider that is potentially biased in the opposite direction, overtake the intertemporal incentives and the full information equilibrium is recovered, even though it is wasteful in terms of social welfare. Hence, the paper provides a model with rational information providers and viewers that leads to biased signals in equilibrium. In the second essay, titled “Voter Behavior and Information Aggregation in Elections with Supermajority”, I provide a model of elections where there are three possible outcomes, but the voters can directly vote for one of the two options. Theoutcome of the election corresponds to the options if the vote share for one of them is higher than a supermajority threshold. If neither of the options achieves that, then the result is the third outcome that the voters cannot explicitly vote for, which I interpret as compromise. I investigate various properties of elections in this setting. I find that, in line with the popular argument, supermajority rules foster compromise outcomes. But, on the other hand, elections with supermajority rules fail to aggregate information. In the third essay, titled “Protests, Strategic Information Provision and Political Communication”, I consider a model of protests where the protesters learn about the state of the world via a biased information provider whose objective is to either instigate or dissuade the protest. A successful protest removes the incumbent from office, where the success threshold is determined by the incumbent who is biased. My main aim is to uncover whether the incumbent can learn the true state of the world from the protest turnout, even though the information of the citizens is provided by biased media. I pin down the optimal success threshold and signal noise choices by the incumbent and the information provider, respectively. I find that if the information provider is trying to instigate the protest, then political communication is always possible, regardless of the level of the bias of the incumbent. If the information provider is trying to dissuade the protest, then political communication is possible if and only if the incumbent bias is relatively small.