Dictatorship has been one of the most persistent regimes types in history. Different dictators have applied different strategies for maintaining political support across different societies. We discuss and empirically estimate the hypothesis that states that dictators rely more on patronage as compared to the general provision of public goods for political support. Our results, based on the data from cross-section of the countries from all continents, confirm this hypothesis. We use military spending as an indicator of the patronage to military and the secondary school enrolment as an indicator of the provision of public goods. In the separate sets of regressions, we conclude that dictatorship has a significant negative effect on the secondary school enrolment rate and a significant positive effect on military expenditure as percentage of GDP. These effects, in turn, might have caused the persistent of dictatorships in many societies. In order to generalise these findings, we also check robustness of the findings with respect to other variables like infant mortality rate, average life expectancy, Human Development Index (HDI), corruption, rule of law, ease of doing business and competitiveness. The robustness analysis confirms our findings.