Social influence on probability judgments and the puzzling stylized facts in financial markets

This dissertation is an exploratory study of financial market implications of social influences on probability judgments. We put forward a preliminary model of social influence on judgments and empirically investigate a conjecture inspired by our simulation results. Our model is an initial tentative attempt at devising an asset pricing model that incorporates social influences on probability judgments.; A decision maker is just a name given to a utility function which is maximized subject to a model, where a model of an agent is his belief regarding the transition law linking the state variable to the control. There is strong empirical evidence that markets have internal dynamics of their own and over the last two decades a series of phenomena that are anomalies under rational expectations finance have been documented. Furthermore, the conclusions from empirical time series literature is that most macroeconomic time series relations are non-stationary with structural instability, hence making it impossible for any agent to have the kind of structural knowledge that the theory demands. Hence, all decision makers face ambiguity. Social psychologists have extensively studied the link between decisions made by an individual under ambiguity and his social context and have documented that social context has a strong influence on an individual's decisions and especially so under ambiguity. This means that economic phenomena under ambiguity are in essence socioeconomic phenomena demanding that we consider both individual economic incentives as well as the social context as the determinant of human behavior. So, a decision maker under ambiguity is a utility function which is maximized subject to a model that is open to social influence.; This dissertation carries out this modification and puts forward a preliminary model of social interactions under ambiguity. The time series generated by the model displays the stylized facts observed in the financial time series that are considered as puzzling or paradoxical under rational expectations based modern finance. The stylized facts generated by our model are volatility clustering and fat tails in returns distribution.; Our model is also consistent with a number of other phenomena observed in financial markets such as asymmetric volatility over a business cycle, the Ramadan effect and time varying kurtosis.