Motivated by the mixed evidence in previous literature, we reexamine the effects of various types of government spending and taxes, as well as overall budget surplus/deficit, on economic growth. To address the model uncertainty issue that may have plagued earlier studies we employ a Bayesian Model Averaging (BMA) approach. We use a panel data set for OECD countries for the 1990-2013 period, control for country and time specific effects, and allow for a wide range of other potential growth determinants. The results suggest a robust link between only some fiscal variables and economic growth. On the spending side, productive public spending has a robust positive effect on growth. On the revenue side, we document a robust negative effect for the top corporate tax rate, but, maybe surprisingly, not for any income tax variable. Finally, our results suggest that a budget surplus has a robust positive effect on economic performance. We also analyze the timing of effects and conclude that most effects occur with a lag of two years.