Semi-presidentialism is the situation where the constitution identifies both a directly elected president and a prime minister responsible to the legislature. There are now some 58 countries in the world with a semi-presidential constitution. However, the academic wisdom is resolutely opposed to the adoption of semi-presidentialism and nascent democracies are advised to avoid this form of government. This paper examines the performance of semi-presidentialism. Particular attention is paid to the effect of various forms of semi-presidentialism. Different forms of semi-presidentialism are expected to have different effects. To this end, all the countries with a semi-presidential constitution that have embarked on the process of democratic transition are identified. To what extent was semi-presidentialism a factor in the cases when the transition process was successful? When the transition process failed, to what extent was semi-presidentialism responsible for this failure? What was the effect of different forms of semi-presidentialism on the process of democratisation? The findings suggest that there is inconclusive evidence to support some of the major problems commonly associated with semi-presidentialism. The performance of semi-presidentialism seems strongly influenced by non-institutional factors. However, there is a difference between the performance of the two main types of semi-presidentialism that are identified. Overall, the findings do not provide grounds to recommend for or against the adoption of semi-presidentialism as opposed to parliamentarism or presidentialism, but if constitution-makers decide to adopt a semi-presidential constitution, then the findings suggest that they should adopt a premier-presidential form of semi-presidentialism.