Alternative models of security price equilibrium
The major determinant of the performance of financial markets is the nature of the information available, both in terms of the overall quality of the information held by investors, and the distribution of information amongst investors. The nature of this issue makes it difficult to model realistically. This thesis marks an attempt to gain insights to the behaviour of securitiesm arkets by investigating the consequences of relaxing, in a realistic way, some of the restrictions on information in existing models. The core of the thesis consists of formal models of the stock market. The first of these is a development of the information aggregation literature, and in particular the model of Hellwig (1980). It looks at the ability of prices to aggregate information that is dispersed among agents who specialise in acquiring information about particular components of the factors that determine the future stock value. We find that narrowing the extent of specialisation beyond a certain point will inevitably lead to a reduction in the informativeness of the market price. The second model is also a development of the information aggregation literature, and looks at the implications of investors obtaining private information about the extent of liquidity trading. We find that such a framework gives rise to the possibility of multiple equilibria and price 'crashes'. The third model is an extension of the second in which the acquisition of information is made endogenousa, nd shows that the main results of that model are retained. It also shows up the dominance of the cost of information about value, rather than about liquidity trading, in determining the overall informativeness of the price. After investigating the possible consequences of, and providing evidence for, the existence of positive feedback trading we investigate the behaviour of a market in which investors exhibiting such behaviour are combined with investors who trade on the basis of value in a form as in De Long, Shleifer, Summers & Waldmann (1989, 1990a). We then apply results from Hart (1977) to determine the conditions under which manipulation can be possible. A number of model characteristics are shown to be possible depending on the specific form of the feedback trading. We finish by adding shocks to the system, as in De Long et. al., and look at the effect of both competitive and monopolistic speculation. We find that competitive speculation may be more destabilising than monopolistic speculation, and that positive feedback trading is more destabilising when it acts after a delay.