This paper empirically revisits the validity of Wagner’s proposition in a panel of 149 developing countries between 1980-2015 by focusing on different components of government expenditure. We rely on an ARDL approach which allow us to uncover short and long-run cyclicality coefficients. Our results do not overwhelmingly support the existence of higher than unity long-run elasticities of government spending components vis-a-vis economic growth, suggesting that the Wagner’s regularity is more the exception than the norm. Moreover, the case for voracity is fading away as developing countries catch-up the development ladder and graduate from procyclicality. In fact, most short-run elasticities are countercyclical. Finally, some macroeconomic and institutional and political characteristics affect the degree of government spending cyclicality.