Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractWe extend the standard approach to Bayesian forecast combination by forming the weights for the model averaged forecast from the predictive likelihood rather than the standard marginal likelihood. The use of predictive measures of fit offers greater protection against in-sample overfitting and improves forecast performance. For the predictive likelihood we show analytically that the forecast weights have good large and small sample properties. This is confirmed in a simulation study and an application to forecasts of the Swedish inflation rate where forecast combination using the predictive likelihood outperforms standard Bayesian model averaging using the marginal likelihood.

Most macroeconomic data are uncertain -they are estimates rather than perfect measures of underlying economic variables. One symptom of that uncertainty is the propensity of statistical agencies to revise their estimates in the light of new information or methodological advances. This paper sets out an approach for extracting the signal from uncertain data. It describes a two-step estimation procedure in which the history of past revisions are first used to estimate the parameters of a measurement equation describing the official published estimates. These parameters are then imposed in a maximum likelihood estimation of a state space model for the macroeconomic variable.Keywords: Real-time data analysis; State space models; Data uncertainty; Data revisions JEL codes: C32, C53 * This paper represents the views and analysis of the authors and should not be thought to represent those of the Bank of England, Monetary Policy Committee, or any other organisation to which the authors are affiliated. We would like to thank the associate editor for extremely helpful comments on an early version of the paper. We have further benefited from helpful comments from Andrew Blake, Spencer Dale, Kevin Lee, Lavan Mahadeva, Tony Yates and Shaun Vahey. We would also like to thank Simon Van Norden for his constructive and incisive comments especially on the relation between our model and comparable recent models in the literature.

Most macroeconomic data are uncertain -they are estimates rather than perfect measures of underlying economic variables. One symptom of that uncertainty is the propensity of statistical agencies to revise their estimates in the light of new information or methodological advances. This paper sets out an approach for extracting the signal from uncertain data. It describes a two-step estimation procedure in which the history of past revisions are first used to estimate the parameters of a measurement equation describing the official published estimates. These parameters are then imposed in a maximum likelihood estimation of a state space model for the macroeconomic variable.Keywords: Real-time data analysis; State space models; Data uncertainty; Data revisions JEL codes: C32, C53 * This paper represents the views and analysis of the authors and should not be thought to represent those of the Bank of England, Monetary Policy Committee, or any other organisation to which the authors are affiliated. We would like to thank the associate editor for extremely helpful comments on an early version of the paper. We have further benefited from helpful comments from Andrew Blake, Spencer Dale, Kevin Lee, Lavan Mahadeva, Tony Yates and Shaun Vahey. We would also like to thank Simon Van Norden for his constructive and incisive comments especially on the relation between our model and comparable recent models in the literature.

We examine how to forecast after a recent break. We consider monitoring for change and then combining forecasts from models that do and do not use data before the change; and robust methods, namely rolling regressions, forecast averaging over different windows and exponentially weighted moving average (EWMA) forecasting. We derive analytical results for the performance of the robust methods relative to a full-sample recursive benchmark. For a location model subject to stochastic breaks the relative MSFE ranking is EWMA < rolling regression < forecast averaging. No clear ranking emerges under deterministic breaks. In Monte Carlo experiments forecast averaging improves performance in many cases with little penalty where there are small or infrequent changes. Similar results emerge when we examine a large number of UK and US macroeconomic series.

We examine how to forecast after a recent break. We consider monitoring for change and then combining forecasts from models that do and do not use data before the change; and robust methods, namely rolling regressions, forecast averaging over different windows and exponentially weighted moving average (EWMA) forecasting. We derive analytical results for the performance of the robust methods relative to a full-sample recursive benchmark. For a location model subject to stochastic breaks the relative MSFE ranking is EWMA < rolling regression < forecast averaging. No clear ranking emerges under deterministic breaks. In Monte Carlo experiments forecast averaging improves performance in many cases with little penalty where there are small or infrequent changes. Similar results emerge when we examine a large number of UK and US macroeconomic series.