I know you wanted to veer back to the topic, but just to add another example to the pile: the math on mass transit is just blindingly obvious. A cheap centralized transit system is wildly more efficient for moving people around than individual cars. And investments in it will "lift all boats" socioeconomically speaking.

But it is more about the problems than the methodology. Centralized/Cooperative methods are more appropriate to solving a complex problem for which you know the solution (building a house, mass transit, maintaining highways). Distributed/Competitive methodology is better for exploring a problem space (determining the most energy efficient house design, innovative bus or train engines, cost effective highway surfaces). If you use cooperative when you should be using competition, you get a cartel. If you use competition when you should be cooperating, you get extreme inefficiency and long delays (I don't think we have a word for when we force groups who should be cooperating to compete with each other for resources).