ABSTRACTIn this paper, we study the relationship between organized crime, corruption, and economic growth on a data set from Italian regions for the period 1996–2013. Our working hypothesis is that organized crime can embezzle part of the public expenditure aimed at productive uses by threatening and bribing public officers. To assess the consequences for regional growth we estimate a finite mixture covariate measurement model and find that the relationship between public expenditure and per capita GDP is characterized by parameter heterogeneity. Specifically, regions are partitioned in clusters identified by the initial level of organized crime. The effect of public expenditure on per capita GDP differs across clusters of regions: in the regions with the higher levels of organized crime public expenditure has a negative effect on per capita GDP, and the estimated share of embezzled public expenditure is higher, amounting to approximately 10% of its book value. Differently, in the regions with lower levels of organized crime the effect of public expenditure on per capita GDP is positive and the estimated share of embezzled public expenditure is lower. The empirical analysis is shown to be consistent with a theoretical growth model à la Barro (1990) augmented by corruption orchestrated by organized crime.