The paper identifies the factors influencing the scope of presidential powers. To test the hypotheses via simple OLS regressions, a new Index of presidential powers has been developed which evaluates presidential powers in comparison with parliamentary powers. The Index is based on the approach of M. Shugart and J. Carey but their way to measure presidential powers was modified in order to be more in line with the aims of the paper. The predictors are: a) the political uncertainty in the period of constitution-drafting process; b) the comparative strength of actors participating in the constitution-drafting; c) index of democracy; d) the breach with the authoritarian past; e) geographic location; f) ethnic fragmentation. The sample consists of countries where a new constitution was adopted after 1970. The results are robust to alternative versions of both independent and dependent variables and to the exception of the most influential observations. The results show that contextual factors (geographic location) are much more significant than factors which imply universal causal mechanisms. The paper indicates that contextual explanations of the scope of presidential powers either can be proposed instead of explanations based on universal causal mechanisms, or they can be used along with the latter and deepen them. The most important limitations of the paper are: small sample size (N = 44) and subjectivity associated with the development of the coding rules which lie behind the Index of presidential powers. As far as the latter is concerned, there is the free access to coding rules via the internet-link in the paper.