Econometric modelling with time series. Specification, estimation and testing (Q2904672)

This book gives a comprehensive introduction into statistical models in econometrics. There is a focus on time series but important econometric models without time series aspect such as classical regression are also treated. Special emphasis is given to maximum likelihood methods which are explained in great detail in Part 1 of the book including estimators, tests and numerical aspects. Part 2 of the book treats regression models starting from linear regression via nonlinear models to regression models with time series errors (including heteroscedastic errors). In Part 3, alternative estimation methods are discussed in detail such as quasi-maximum-likelihood, generalized method of moments, nonparametric estimation (in particular kernel methods) as well as estimation by simulation. The remaining three parts deal with time series models where multivariate models are included right from the start. Part 4 introduces stationary models, from linear time series via structural vector autoregressions to latent factor models. Part 5 deals with non-stationary models such as integrated and cointegrated time series. Finally, Part 6 discusses nonlinear time series, where nonlinearities in the mean as well as variance are considered, and a full chapter is devoted to discrete time series.NEWLINENEWLINEThe authors focus on examples and statistical methodology, where the examples and topics discussed include very recent material. While important mathematical results are usually given, their proofs are left to other books but corresponding references are often missing. The computer code used in the examples can be downloaded from a companion website in three different languages (R, Matlab and Gauss).NEWLINENEWLINEIn conclusion, this book is very suitable for readers who are looking for examples for both lectures and research as well as readers mainly interested in statistical methodology. For lectures aimed mainly at mathematics students (including Time Series Analysis, Regression Models and Nonparametric Statistics) it can be used as additional material giving useful examples and providing relevant code but is not suitable as a main text due to the missing mathematical details, which can however be found in different books.