| My dissertation is written in the fields of finance and econometrics. The research agenda starts with a Monte Carlo study, one which reveals structural limitations of the standard square-root model in fitting the term structure of interest rates and the finite sample inefficiency of simulation-based estimators in empirical finance. The discovery of these limitations naturally leads to the two major contributions of my thesis research. One is to develop new term structure models---ones that have discrete jumps or regime shifts---that are in line with the general equilibrium no-arbitrage pricing approach. The other significant contribution is to construct highly efficient econometric methods to estimate jump-diffusion and stochastic volatility processes, by providing closed form solutions to the conditional moments of the underlying point-in-time process or the integrated time series. The major findings of the research are these: the jump-augmented or regime-augmented yield curves are much more flexible in fitting the observed term structure, and the conditional-moment-based econometric estimators are statistically reliable and computationally efficient. What is gained from these new modeling and estimation strategies is that one may adequately capture the rich volatility pattern embedded in most financial time series data. |
