| In the first chapter, we examine the role that beliefs, prices and participation externalities play in market collapses and show that pessimistic beliefs can cause a market collapse even when the state is good. After receiving consecutive negative shocks, some ex-ante identical Bayesian agents will be discouraged about the unknown state of the economy; therefore, they will exit the market. This will decrease prices due to the diminished participation externality. The agents who remain in the market learn the true state from this change in prices. If these fully-informed agents are not too many, then they will exit, causing an inefficient market collapse.; In the second chapter, we study capital outflows and unemployment caused by it. At the domestic country, capital owners have access to a long-term production technology which uses labor and capital. Labor and capital agree how to share the uncertain return of the long term technology at the beginning. At each period, Bayesian capital owners get signals and update their prior beliefs about the riskiness of the return. When some of them believe that investing in world markets is better than using the domestic technology, an initial capital outflow will occur. This initial capital outflow will make everyone fully informed. The fully informed workers might agree to be paid less then they were promised in order to stop a second round of capital outflow and prevent being unemployed.; In the third chapter, in an evolutionary game theory setting, we deal with the equilibrium selection issue in a coordination game when there is tension between the risk dominant equilibrium and the Pareto dominant equilibrium. By using heterogenous locations, moving costs and congestion effects, we show that if there are fewer agents in one location when the risk dominant actions are played in both locations, then it is easier for the agents to learn to coordinate on the Pareto dominant equilibria in both locations. Also, it is shown that the lower the moving costs, the more likely to have the Pareto dominant equilibria in both locations. We, also, consider the N locations case. |
