Time Series Models

This volume consists of the revised versions of the main papers given at the second Seminaire Europeen de Statistique on 'Likelihood, Time Series, with Econometrics and Other Applications', held at Nuffield College, Oxford from 13-17 December 1994. The aim ofthe Seminaire Europeen de Statistique is to provide talented young researchers with an opportunity to get quickly to the forefront of knowledge and research in areas of current major focus. Accordingly, as in the book based on the first seminar in the series, 'Networks and Chaos - Statistical and Probabilistic Aspects', the papers in this volume have a tutorial character. In the present Seminaire about 35 young statisticians from ten European countries participated. Nearly all participants gave short presentations about their recent work; these, while of high quality, are not reproduced here. The paper by N. G. Shephard reviews and extends work on a class of nonlinear time series models widely used in econometrics and of potential interest in other fields. S. Johansen gives a widely accessible account of cointegration, an important notion in the interpretation of multivariate nonstationary time series. M. P. Clements and D. F. Hendry give a general discussion of the statistics of forecasting errors. These three papers draw their motivation directly from econometrics. By contrast, N. Laird discusses methods developed in a biostatistical context for the analysis of short time series. Finally, B. A. Jensen and J. A.