The Eurosystem/ECB staff macroeconomic projection exercises constitute an important input to the ECB's monetary policy. This work marks a thorough analysis of the Eurosystem/ECB projection errors by looking at criteria of optimality and rationality using techniques widely employed in the applied literature of forecast evaluation. In general, the results are encouraging and suggest that Eurosystem/ECB staff projections abide to the main characteristics that constitute them reliable as a policy input. Projections of GDP - up to one year - and inflation are optimal - in the case of inflation they are also rational. A main finding is that GDP forecasts can be substantially improved, especially at long horizons.