Document Type
Conference Item
Publication Date
1-1-2010
Abstract
Autoregressive Conditional Duration (ACD) models playa central role in modelling high frequency financial data. The Maximum Likelihood (ML) and Quasi Maximum Likelihood (QML) methods are widely used in parameter estimation. This paper considers a semi parametric approach based on the theory of Estimating Function (EF) in estimation of A CD models. We use a number of popular distributions with positive supports for errors and estimate the parameter(s) using the both EF and . ML approaches. A simulation study is conducted to compare the performance of the EF and the corresponding ML estimates for ACD(1.1), ACD(l,2) and ACD(2,l) models. It is shown that the EF approach provides comparable estimates with the ML estimates using a shorter computation time. Finally, both methods are applied to model a real financial data set and provide empirical evidence to support the use EF approach in practice.
Keywords
Conditional duration, Estimating function, High Frequency data Maximum likelihood
Event Title
International Statistics Conference 2010
Event Location
Colombo, Sri Lanka
Event Dates
08-09 Jan 2010
Event Type
conference