<p><p>We survey the recent empirical literature concerning the cyclical response of fiscal policies in the euro area, finding large differences in results. We show that these differences are heavily influenced by the choices made in modelling fiscal behaviour. We make a case for the use, in assessing policies and in the EMU context, of the standard modelling choice where the discretionary reaction of fiscal policy is directly estimated. Models where the overall reaction to the cycle - which includes the effects of both discretionary actions and automatic stabilisers - is directly estimated tend to suggest either strong pro-cyclical or strong counter-cyclical discretionary reactions, depending on how this component is identified. Within the standard model and for the years 1994 to 2008, we show that results are significantly affected by the data vintage (ex post or real-time). With ex post data, we find an unambiguous indication of a-cyclicality. Using real-time data, we find that the output gap matters. However, depending on whether we assess policy intentions or actual policies, euro-area governments' behaviour radically changes. A plausible interpretation is that in the implementation phase, governments loosen their fiscal stance, giving in to political pressures that are proportional to the scale of the economic difficulties in bad times and the size of the 'growth dividend' in good times.</p></p>