Comparing Estimation Procedures For Stochastic Volatility Models Of Short Term Interest Rates

The Stochastic Volatility of Short-term Interest Rates

Comparing Estimation Procedures for Stochastic Volatility Models of Short-Term Interest Rates

This paper compares the performance of three maximum likelihood estimation procedures -quasi-maximum likelihood, Monte Carlo likelihood and the particle filter to estimate stochastic volatility models of short term interest rates. The procedures are compared in an empirical study of interest rate volatility where a number of diagnostic tests in- and out-of-sample are utilized to evaluate both model specification and estimation procedure. Empirically, the results suggest interest rates follow the Cox-Ingersoll-Ross model with stochastic volatility and that volatility increases after Federal Open Market Committee meetings. Overall, the Monte Carlo likelihood procedure provided the best results.

Stochastic Filtering With Applications In Finance

This book provides a comprehensive account of stochastic filtering as a modeling tool in finance and economics. It aims to present this very important tool with a view to making it more popular among researchers in the disciplines of finance and economics. It is not intended to give a complete mathematical treatment of different stochastic filtering approaches, but rather to describe them in simple terms and illustrate their application with real historical data for problems normally encountered in these disciplines. Beyond laying out the steps to be implemented, the steps are demonstrated in the context of different market segments. Although no prior knowledge in this area is required, the reader is expected to have knowledge of probability theory as well as a general mathematical aptitude.Its simple presentation of complex algorithms required to solve modeling problems in increasingly sophisticated financial markets makes this book particularly valuable as a reference for graduate students and researchers interested in the field. Furthermore, it analyses the model estimation results in the context of the market and contrasts these with contemporary research publications. It is also suitable for use as a text for graduate level courses on stochastic modeling.