Estimating time-variation in measurement error from data revisions; an application to forecasting in dynamic models

Working papers set out research in progress by our staff, with the aim of encouraging comments and debate.
Published on 12 November 2004

Working Paper No. 238
By George Kapetanios and Tony Yates

Over time, economic statistics are refined. This means that newer data are typically less well measured than old data. Time or vintage-variation in measurement error like this influences how forecasts should be made. Measurement error is obviously not directly observable. This paper shows that modelling the behaviour of the statistics agency generates an estimate of this time-variation. This provides an alternative to assuming that the final releases of variables are true. The paper applies the method to UK aggregate expenditure data, and demonstrates the gains in forecasting from exploiting these model-based estimates of measurement error.

PDFEstimating time-variation in measurement error from data revisions; an application to forecasting in dynamic models

Other papers

Was this page useful?
Add your details...