There are certain scenarios that work, where the incentives are properly aligned in time, like the Invisible Hand scenario described by Adam Smith.

When it is possible to steal from the future, however, it's another thing. The collective action problem linked there is about the problems a group has acting in its own interest when control is too atomized. A group like

https://en.wikipedia.org/wiki/Koch_network

is a political activist organization because the people who care about the results can defund it if it veers from its goal, whereas membership organizations like

https://en.wikipedia.org/wiki/Greenpeace

or

https://en.wikipedia.org/wiki/Sierra_Club

ultimately become organizations that steal from their members because of the problems described here:

https://en.wikipedia.org/wiki/Exit,_Voice,_and_Loyalty

If an apartment building is owned by a rich individual the government can make them put up the cash or get a loan or raise the rent (some combination of which, one way or another private capital provides degrees of freedom that can part of the answer. In the case of the condo the individuals don't have the money, they can't get a loan, they can't sell their shares in a failing building, etc.

Those are all fine examples but we see cooperative action on both smal and large scales, formal and informal... the heap of edge cases becomes so large i=t becomes clear that Hume is just not correct. It's more complicated.