OPTIMAL LINEAR INVESTMENT STRATEGIES FOR A STACKLEBERG LEADER IN A RATIONAL EXPECTATIONS MODEL OF A SPECULATIVE MARKET

This thesis analyzes non-price-taking behavior and its implications for prices in a speculative market. A two period model is developed in which there are price-taking speculators and a non-price-taking speculator, each with a linear demand function for a risky asset and some private information about the "true value" of the asset. Each speculator chooses the parameters of this linear demand function so as to maximize some objective function using the private information that he possesses and the information of other speculators that is communicated to him by the equilibrium price of the risky asset. A non-fully-informing linear rational expectations equilibrium is shown to exist and a number of parameter restrictions are derived from the model's assumptions. This has both normative and positive implications. On the normative side, parameter restrictions are imposed on the demand functions that are associated with the equilibrium. These describe optimal linear investment strategies. In addition, the restrictions on the equilibrium price function and on the optimal demands place bounds on the expected profit and expected utility of investors, and these bounds vary depending on the regulatory environment. This allows a regulatory authority to order various regulatory schemes according to the preferences of investors. On the positive side, this model gives a description of the workings of a securities market. We obtain an understanding of why this description sometimes conforms and sometimes doesn't conform to previously held beliefs about such markets. In addition to the usual comparative statics results, a number of well known empirical regularities are given different interpretations because the model is quite different from the established capital market models in certain structural areas. Perhaps, most importantly, the model shows that non-price-taking behavior in a stochastic environment can be modeled in a simple equilibrium framework and that an empirically testable comparison with price-taking behavior is possible in the same framework. This suggests an important direction for future theoretical and empirical research.