| This thesis is comprised of three essays which discuss issues in financial econometrics from a Bayesian perspective. The first essay, "Inference in Simultaneous Equations Models with Limited Dependent Variables: An Application to the Study of Default", surveys the sampling theoretic approach and develops new inferential procedures for conducting a finite sample likelihood-based analysis of simultaneous equations models with limited dependent variables. The merits of these new methods are shown to be worth the effort. The proposed techniques are then applied to default data of original issue high yield bonds.;The second essay, "Exchange Rate Target Zone Models: A Bayesian Evaluation", utilizes the techniques developed in the first essay. It presents a uniform Bayesian approach to estimation, prediction and model comparison for exchange rate target zone models. It proposes a simultaneous-equations target zone model with stochastic devaluation risk. This model is compared with the existing alternatives. Two empirical examples illustrate the procedures and our findings.;The third essay, "Bayesian Analysis of Duration Models: An Application to Bankruptcy", develops an estimable model of instantaneous probability (hazard rate) of a firm's exit from formal reorganization. Using Bayesian methodology, the effects on the hazard rate of various firm-specific financial characteristics and of the time spent under Chapter 11 are analysed. |
