Recent research on growth empirics has focused on resolving model and variable uncertainty. The conventional approach has been to assume a linear growth process and then to proceed with investigating the relevant variables that determine cross-country growth. This article questions the linearity assumption underlying the vast majority of such research and uses recently developed non-parametric techniques to handle non-linearities as well as select relevant variables. We show that inclusion of non-linearities is necessary for determining the empirically relevant variables and uncovering key mechanisms of the growth process.