General - to - Specyfic Modeling 2 vols

'As data sets have become larger, computers greatly increased in capacity and speed, exploratory techniques have grown more sophisticated and models are now multivariate, dynamic, and possibly nonlinear. The number of models that could be considered, and thus compared, have become immense. Econometricians need to be helped through this tangled maze and this help can be found in these volumes. Here leading advocates of a variety of well-tried and appreciated approaches to modeling display their expert knowledge.' - Sir Clive W.J. Granger, University of California, San Diego, US and 2003 Nobel Laureate in Economics Economists have long sought to develop quantitative models of economic behaviour, which blend economic theory with data evidence. Econometric modelling of economic time series has strived to achieve this by seeking to discover sustainable and interpretable relationships. This important two-volume collection focuses on a central method used in selecting such models, namely simplification of an initially general model that adequately characterizes the empirical evidence within the investigators' theoretical framework. The volumes feature a wealth of evidence that has accrued over the last five years displaying its excellent abilities for model selection, based on Monte Carlo studies of automatic algorithms. These also throw light on several major methodological issues, and prompt many new ideas, which are discussed. The collection will be valuable to all empirical economists and econometricians.