Are institutions a deep cause of economic growth? This paper tries to answer this question in a novel manner by focusing on within-country variation, over long periods of time, using a new hand-collected data set on institutions and the power-ARCH econometric framework. Focusing on the case of Brazil since 1870, our results suggest (a) that both changes in formal political institutions and informal political instability affect economic growth negatively, (b) there are important differences in terms of their short- versus long-run behaviour, and (c) not all but just a few selected institutions affect economic growth in the long-run.
