Real-time consideration of systematic patterns in Maltese GDP data and revisions

Economic policies are generally formulated on the basis of data available in real time, which might subsequently be revised. Implicitly, the possibility of data revisions creates an element of uncertainty around the very same data driving policy decisions. Given that such uncertainty could be mitigated if data revisions were predictable, this paper sets out to identify any systematic patterns in the Maltese 'Real GDP' data and underlying revisions which could in turn prove useful when making economic assessments in real time. These patterns are identified through the application of VAR econometric techniques applied in previous economic literature to model Malta's first-release output data alongside the revisions made within a maximum of two years from the date of first release. Systematic patterns in Maltese data are recorded, and in turn used to anticipate in real time the revisions that might be made upon the publication of future vintages. The analyses conducted in this paper show that the real-time consideration of these forecasts has the potential to improve the precision with which past, contemporaneous and (forecasted) future Maltese output data could be interpreted in real time. The uncertainty surrounding these forecasts is also aptly demonstrated through the use of simulation techniques, density functions, and event probability forecasts.