Although semi-presidentialism is a popular form of governance in new democracies, we have little empirical evidence supporting its popularity. In this study, I attempt to reassess democratic performance of new semi-presidential regimes from 1974 to 2009 as a function of four broadly cited risk factors: divided minority government, cohabitation, presidential powers and president-parliamentarism. The results are more encouraging than previous research has suggested. First, divided minority government is positively associated with higher levels of democracy, even though it, along with a strong presidency and president-parliamentarism, makes executive instability more likely. Second, perils of cohabitation are not substantiated with regard to executive instability and quality of democracy. Third, none of the caveats against semi-presidential systems makes them more vulnerable to democratic breakdown. However, a failure to check presidential powers appears to be a serious risk for semi-presidentialism. As presidents enjoy more powers, the levels of democracy tend to decrease. This finding has a substantive implication for countries that already practise semi-presidential governance or contemplate a move in that direction: checking presidential powers is critical to facilitate democratic consolidation in semi-presidentialism.