Using a novel panel dataset of 65 countries for the period 1973-2000, this paper attempts to reconcile the conflicting evidence provided by the current literature on the effects of different tax categories on long-term economic growth. The effects of both top and average, personal and corporate income tax rates as well as the level of tax progressivity on growth are tested while controlling for other possible determinants of growth. The empirical results provide evidence for the distortionary effects of personal income taxes and tax progressivity on long run economic growth, but no robust evidence for any linear effect of corporate tax rates on long term economic growth is found. There is however, evidence for a non-linear effect of corporate taxation on economic performance. The sample splitting estimations also yield thresholds above which progressivity becomes harmful for growth but below which there is no significant effect. The thresholds associated with the respective decades seem to follow the average degree of progressivity rather closely and could be indicative of international tax competition.