We estimate fiscal multipliers in a panel of countries using dynamic panel techniques and quarterly data for 55 countries. By using a GMM estimator and lagged dependent variables as instruments in a SVAR model, we attempt to correct for the biases present in this setting, to alleviate concerns about causality, and to decrease potential effects of third factors. Contrary to previous research, we find no strong evidence of monetary accommodation, a positive and larger fiscal multiplier in developing than in high-income countries, and zero in high-debt countries and in flexible exchange rates countries.