Finance Discipline Group
UTS Business School
University of Technology, Sydney

Working Paper Series

Title:
Continuous Time Model Estimation
Author(s): Carl Chiarella & Shenhuai Gao
Date of publication: December 2004
Working paper number: 138
Abstract:
This paper introduces an easy to follow method for continuous time model estimation. It serves as an introduction on how to convert a state space model from continuous time to discrete time, how to decompose a hybrid stochastic model into a trend model plus a noise model, how to estimate the trend model by simulation, and how to calculate standard errors from estimation of the noise model. It also discusses the numerical difficulties involved in discrete time models that bring about the unit roots illusion in econometrics.
Paper: Download (Format: PDF, Size: 155 Kb)