Time-series modelling in financial markets : new approaches and exchange rate applications

This thesis investigates and develops time-series models in financial markets. It provides new approaches to this area, and applies these approaches to various exchange rate contexts. The aims of this thesis are: to develop an understanding of modelling international financial market movements, in particular the foreign exchange market and the Australian equity market; to examine advances in non­ linear time-series modelling using computer-intensive statistical approaches; and to investigate and develop non-linear time-series modelling in financial markets. The topic of time-series modelling in financial markets has received widespread coverage in the literature. The issue is important because of its fundamental role in investment decision-making. Over the last decade, financial markets have been affected by significant structural changes, including the expansion of financial markets, the globalisation of finance, the introduction of the Euro, the increasing role of electronic broking, and the massive growth in the investment funds industry. These factors have contributed to an increased focus on financial market movements. Conventional linear time-series models are unlikely to properly explain and model significant price variations. Non-linear time-series models have the potential to improve the performance of modelling and are appealing in explaining complex price behavior.