Research suggests that institutions affect the levels of corruption in a country. We take these arguments a step further and examine whether it is the presence of inclusive institutions and/or the credible and consistent implementation of institutions that matter, as regards corruption. We use a novel approach to theoretically conceptualise and empirically operationalise institutions along two analytically distinct dimensions: the nature of the institutions (the de jure dimension), and the extent to which they are credibly and consistently implemented over time (...) (the de facto dimension), using a panel dataset for 148 countries covering 2012 to 2018. We find that formal institutions are most effective in reducing corruption when the rules are credibly and consistently implemented. Furthermore, this effect appears to operate differently at different levels of national income. The nature of the formal institutions appears important across income levels, but particularly so at the upper middle income level, while the credible and consistent implementation of these institutions is primarily influential at upper middle income levels. We explain why this may be the case and elaborate on the policy implications. (shrink)