Public-private partnerships (PPPs) have become an essential vehicle for infrastructure development worldwide. Theoretical arguments primarily focus on buildoperate-transfer (BOT) agreements as a canonical form of PPP, though they rarely discuss the political underpinnings of governments' decisions to enter such agreements. How does a government's longevity, stability and its capacity to raise revenue make BOTs more attractive than other types of partnerships? Extending recent theoretical advances through concepts of control rights and veto players and statistically analyzing a database of more than 4,300 PPP agreements for new construction of infrastructure in 83 developing economies between 1990-2014, I provide the first large-scale quantitative evidence of the influence of political institutions on government choices to adopt BOTs. I find that BOTs are less attractive as the tenure of the longest-serving veto player increases, when veto players are more frequently replaced, and when governments can generate more tax revenue, but more likely when that revenue is above a country's historic average. My findings contribute to literatures on distributive public policy, hybrid governance, complex project management, and to the policy debate about the role of PPPs in economic development.