Dynamic Econometrics

Dynamic Econometrics presents a systematic and operational approach to econometric modelling, based on the outcome of a twenty-year research programme. It addresses the practical difficulties of modelling data when the mechanism is unknown, with theory and evidence interlinked at every stage of the discussion. The main problem in econometric modelling of time series is discovering sustainable and interpretable relationships between observed economic variables. This book develops an econometric approach which sustains constructive modelling, clarifies the status of empirical econometric models, and formulates structured tools for critically appraising evidence. Professor Hendry deals with methodological issues of model discovery, data mining, and progressive research strategies, and with major tools for modelling (including recursive methods, encompassing, super exogeneity, and invariance tests). In addition, he considers practical problems of collinearity, heteroscedacity, and measurement errors, and includes an extensive study of UK money demand. The book is self contained, with technical background covered in appendices of matrix algebra, probability theory, regression, asymptotic distribution theory, numerical optimization, and macro-econometrics. Mathematical results appear in solved examples and exercises, and live classroom teaching of econometrics via computer demonstrations is stressed. The structure of the book makes it of practical value to economists investigating empirical phenomena, to advanced undergraduate and graduate econometrics students, and to statisticians involved in the analysis of social science time series.