There is evidence in the literature of fiscal consolidation episodes producing (non-Keynesian) expansionary effects (e.g. Alesina and Ardagna, 1998). We replicate this result for a panel of OECD countries under exogeneity of the fiscal tightening decision, and provide evidence that this decision is endogenous to GDP so that the exogeneity assumption might be inappropriate. Once this endogeneity is taken into consideration, we find that fiscal consolidations have a negative impact on GDP as expected in a Keynesian framework. We also investigate the determinants of successful consolidations. In particular, we use model averaging to overcome the problem of model uncertainty, and conclude that economic recovery and cuts in public wages are the most important ingredients of a consolidation program for successfully reducing budget deficits.