This edited collection concerns nonlinear economic relations that involve time. It is divided into four broad themes that all reflect the work and methodology of Professor Timo TerAcsvirta, one of the leading scholars in the field of nonlinear time series econometrics. The themes are: Testing for linearity and functional form, specification testing and estimation of nonlinear time series models in the form of smooth transition models, model selection and econometric methodology, and finally applications within the area of financial econometrics. All these research fields include contributions that represent state of the art in econometrics such as testing for neglected nonlinearity in neural network models, time-varying GARCH and smooth transition models, STAR models and common factors in volatility modeling, semi-automatic general to specific model selection for nonlinear dynamic models, high-dimensional data analysis for parametric and semi-parametric regression models with dependent data, commodity price modeling, financial analysts earnings forecasts based on asymmetric loss function, local Gaussian correlation and dependence for asymmetric return dependence, and the use of bootstrap aggregation to improve forecast accuracy. Each chapter represents original scholarly work, and reflects the intellectual impact that Timo TerAcsvirta has had and will continue to have, on the profession.The Dynamic Conditional Beta model is able to uncover the changing structure of returns in commodity markets. The volatilities and correlations change over time as the economic structure of the asset class changes. Application of asymmetricanbsp;...
|Title||:||Essays in Nonlinear Time Series Econometrics|
|Author||:||Niels Haldrup, Mika Meitz, Pentti Saikkonen|
|Publisher||:||OUP Oxford - 2014-06-26|