This dissertation is composed of four essays. First, I introduce a method of choice under uncertainty that combines a dynamic optimization procedure with a rule of thumb for expectations formation. This assumption for expectations formation is similar to the assumption, in stochastic models, that the agent\u27s expectations are based on the random walk model. The method of optimal control—neighboring optimal control—formulates optimal stabilizing decisions. This method is then applied to a game of interaction between a representative foreign investor and a host government. I study the tax policy of an opportunistic government. The government is concerned with maximizing the level of revenue and with minimizing fluctuations in revenue. The p...