The Oxford handbook of economic forecasting

The Oxford handbook of economic forecasting, edited by Michael P. Clements and David F. Hendry, Oxford, Oxford University Press, 2011, xv+712 pp., £95.00 or US$150.00 (hardback), ISBN 978-0-19-539864-9

The oxford handbook of economic forecasting follows the paradigm of Elliott et al. 1 and indeed, as stated on its cover, it ‘provides an up-to-date coverage of both new and well-established fields in the sphere of economic forecasting’. The chapters are written by world experts in their respective fields and despite the large number of contributors, the flow is homogeneous as if the volume was written by a single author. The handbook is divided into six parts: the first four focus on methodological issues whereas the last two discuss specific applications.

Part I covers models and methods including vector autoregressive models, dynamic factor models, fully parametric nonlinear specifications (like smooth transition regressions) and unobserved component models; the above-mentioned topics were also covered in 1. The next two parts discuss issues that have attracted considerable interest during the last decade: mixed-frequency data combination methods, forecasting structural breaks and forecasting during breaks and forecast combination methods with regime-specific weights. Part IV is comprised of five chapters on forecast evaluation methods; in this part, the reader will find recent developments in statistical tests of predictive ability.

The two chapters of Part V are devoted to financial forecasting with specific focus on volatility forecasting. Finally, the last part discusses applications that are not directly related to economic forecasting such as weather and election forecasting. Applications are not confined to the last parts of the handbook; typically, in every chapter, methods are illustrated with detailed implementations and occasional references to available computational tools. It is perhaps strange that neither regional economic forecasting using space–time models, nor models for network flows are included in the applications.

It should be clarified that the volume focuses on fully parametric model specifications; forecasting using neural networks or support vector machines is not discussed, despite their relatively frequent use in economic forecasting during the past decade (see e.g. 2). No book can cover all topics, especially in such a popular field; in my opinion, the handbook is very well written and achieves its aim which is to provide ‘authoritative yet accessible accounts of the key concepts, subject matter, and techniques in a number of diverse but related areas’. The volume with its wide variety of methodological topics and applications is definitely useful to practitioners, including the ones that tend to approach forecasting based on machine-learning techniques.