We analyze the forces that explain inflation using a panel of 122 countries from 1997 to 2015 with 37 regressors. 98 models motivated by economic theory are compared to a gradient boosting algorithm, non-linearities and structural breaks are considered. We show that the typical estimation methods are likely to lead to fallacious policy conclusions which motivates the use of a new approach that we propose in this paper. The boosting algorithm outperforms theory-based models. We confirm that energy prices are important but what really matters for inflation is their non-linear interplay with energy rents. Demographic developments also make a difference. Globalization and technology, public debt, central bank independence and political characteristics are less relevant. GDP per capita is more relevant than the output gap, credit growth more than M2 growth.