ABSTRACTWe apply graphical modelling (GM) theory to identify fiscal policy shocks in SVAR models of the US economy. Unlike other econometric approaches – which achieve identification by relying on potentially contentious a priori assumptions – GM is a data based tool. Our results are in line with Keynesian theoretical models, being also quantitatively similar to those obtained in the recent SVAR literature à laBlanchard and Perotti (2002), and contrast with neoclassical real business cycle predictions. Stability checks confirm that our findings are not driven by sample selection.