The first releases of official statistics are often revised in subsequent releases, sometimes
substantially. Such revisions can impact on policy decisions, as revisions to first published data may alter
the previous assessment of the state of the economy. This may occur through a changed interpretation
based on the revised data itself or the impact the revision may have on econometric models which may
incorporate several statistics, each subject to revision. Whilst this is a recognised issue of key importance,
most producers of official statistics do not quantify expected revisions to their data and economists do not
have the required data to test the sensitivity of their econometric models to revisions in input data. This
important gap in knowledge required to effectively use official statistics, and demands from central banks
motivated the OECD to develop a unique new product: the Main Economic Indicators Original Release
Data and Revisions Database, now freely available at: http://stats.oecd.org/mei/default.asp?rev=1 .
Accessing this source of originally published data will enable analysts to test the likely effectiveness of
econometric models in simulated real-time. It will also enable producers of official statistics to study the
magnitude and direction of subsequent revisions to published data which can lead to a better understanding
of the statistical compilation process, enabling problems to be identified and improvements to be made.
Revisions analysis also provides important information to users on the robustness of first estimates...
