A casual glance at the relevant literature suggests that the amount of nonlinear time series models that can be potentially useful for modelling and forecasting economic time series is enormous. Practitioners facing this plethora of models may have difficulty choosing the model that is most appropriate for their particular application, as very few systematic accounts of the pros and cons of the different models are available. In this book we provide an in-depth treatment of several recently developed models, such as regime-switching models and artificial neural networks. We narrow our focus to examining their potential applicability for describing and forecasting financial asset returns and their associated volatilities. The models are presented in substantial detail and are not treated as ‘black boxes’. All models are illustrated on data concerning stock markets and exchange rates.

Our book can be used as a textbook for (advanced) undergraduate and graduate students. In fact, this book emerges from our own lecture notes prepared for courses given at the Econometric Institute, Rotterdam and the Tinbergen Institute graduate school. It must be stressed, though, that students must have had a solid training in mathematics and econometrics and should be familiar with at least the basics of time series analysis. We do review some major concepts in time series analysis in the relevant chapters, but this can hardly be viewed as a complete introduction to the field.