This paper compares the forecasting performance of different models which have been proposed for forecasting in the presence of structural breaks. These models differ in their treatment of the break process, the model which applies in each regime and the out-of-sample probability of a break occurring. In an extensive empirical evaluation involving many important macroeconomic time series, we demonstrate the presence of structural breaks and their importance for forecasting in the vast majority of cases. We find no single forecasting model consistently works best in the presence of structural breaks. In many cases, the formal modeling of the break process is important in achieving good forecast performance. However, there are also many cases where simple, rolling window based forecasts perform well.

Structural change is a fundamental concept in economic model building. Statistics and econometrics provide the tools for identification of change, for estimating the onset of a change, for assessing its extent and relevance. Statistics and econometrics also have de veloped models that are suitable for picturing the data-generating process in the presence of structural change by assimilating the changes or due to the robustness to its presence. Important subjects in this context are forecasting methods. The need for such methods became obvious when, as a consequence of the oil price shock, the results of empirical analyses suddenly seemed to be much less reliable than before. Nowadays, economists agree that models with fixed structure that picture reality over longer periods are illusions. An example for less dramatic causes than the oil price shock with similarly profound effects is economic growth and its impacts on the economic system. Indeed, economic growth was a motivating concept for this volume. In 1983, the International Institute for Applied Systems Analysis (IIASA) in Laxen burg/ Austria initiated an ambitious project on "Economic Growth and Structural Change".

This Handbook provides 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 provide authoritative yet accessible accounts of the key concepts, subject matter, and techniques in a number of diverse but related areas. It covers the ways in which the availability of ever more plentiful data and computational power have been used in forecasting, in terms of the frequency of observations, the number of variables, and the use of multiple data vintages. Greater data availability has been coupled with developments in statistical theory and economic analysis to allow more elaborate and complicated models to be entertained; the volume provides explanations and critiques of these developments. These include factor models, DSGE models, restricted vector autoregressions, and non-linear models, as well as models for handling data observed at mixed frequencies, high-frequency data, multiple data vintages, methods for forecasting when there are structural breaks, and how breaks might be forecast. Also covered are areas which are less commonly associated with economic forecasting, such as climate change, health economics, long-horizon growth forecasting, and political elections. Econometric forecasting has important contributions to make in these areas along with how their developments inform the mainstream.

This book provides a formal analysis of the models, procedures, and measures of economic forecasting with a view to improving forecasting practice. David Hendry and Michael Clements base the analyses on assumptions pertinent to the economies to be forecast, viz. a non-constant, evolving economic system, and econometric models whose form and structure are unknown a priori. The authors find that conclusions which can be established formally for constant-parameter stationary processes and correctly-specified models often do not hold when unrealistic assumptions are relaxed. Despite the difficulty of proceeding formally when models are mis-specified in unknown ways for non-stationary processes that are subject to structural breaks, Hendry and Clements show that significant insights can be gleaned. For example, a formal taxonomy of forecasting errors can be developed, the role of causal information clarified, intercept corrections re-established as a method for achieving robustness against forms of structural change, and measures of forecast accuracy re-interpreted.

In 1984, the University of Bonn (FRG) and the International Institute for Applied System Analysis (IIASA) in Laxenburg (Austria), created a joint research group to analyze the relationship between economic growth and structural change. The research team was to examine the commodity composition as well as the size and direction of commodity and credit flows among countries and regions. Krelle (1988) reports on the results of this "Bonn-IIASA" research project. At the same time, an informal IIASA Working Group was initiated to deal with prob lems of the statistical analysis of economic data in the context of structural change: What tools do we have to identify nonconstancy of model parameters? What type of models are particularly applicable to nonconstant structure? How is forecasting affected by the presence of nonconstant structure? What problems should be anticipated in applying these tools and models? Some 50 experts, mainly statisticians or econometricians from about 15 countries, came together in Lodz, Poland (May 1985); Berlin, GDR (June 1986); and Sulejov, Poland (September 1986) to present and discuss their findings. This volume contains a selected set of those conference contributions as well as several specially invited chapters.

Econometric models are made up of assumptions which never exactly match reality. Among the most contested ones is the requirement that the coefficients of an econometric model remain stable over time. Recent years have therefore seen numerous attempts to test for it or to model possible structural change when it can no longer be ignored. This collection of papers from Empirical Economics mirrors part of this development. The point of departure of most studies in this volume is the standard linear regression model Yt = x;fJt + U (t = I, ... , 1), t where notation is obvious and where the index t emphasises the fact that structural change is mostly discussed and encountered in a time series context. It is much less of a problem for cross section data, although many tests apply there as well. The null hypothesis of most tests for structural change is that fJt = fJo for all t, i.e. that the same regression applies to all time periods in the sample and that the disturbances u are well behaved. The well known Chow test for instance assumes t that there is a single structural shift at a known point in time, i.e. that fJt = fJo (t