The usefulness and relevance of New Public Management (npm) reforms for developing/transitional countries have been widely debated. Three arguments in particular have crossed-cut the literature. First, an argument that might be called "reform by stages", which suggests the need to building a bureaucratic structure à la Weber before attempting neo-managerial reforms. The second one, which might be called "reform by leapfrogging", implies the opposite: the possibility of implementing npm even if a "traditional" public administration has not been developed yet. Lastly, the third one proposes the possibility of designing reforms by way of an "informed combination", that is by mixing components from different models or paradigms. This essay seeks to analyze one by one these arguments, in order to show that each one makes incisive and important assertions, but is also based on disputable and even wrong assumptions.