There's plenty of things that happen that make many people worse off, with no visible compensation. Justice isn't a real thing. This applies across time for a single person as well: decisions you make now can be pleasant or unpleasant for near-term you and the opposite for future-you. There's no utility or resource flow that makes this happen (unless you include empathy and imagination, but this kind of trade framework is very hard to use with such concepts).
You should read Richard Epstien's _Takings_ https://www.amazon.com/Takings-Private-Property-Eminent-Domain/dp/0674867297
It's all about this. He makes a lot of insightful points - we could be improving things far more than we do now, if only we could pay the losers to stop opposing the changes.
The world has lots of people and things in it. And they are organized in such a mishmash that changing a thing will often make large numbers of people better off or worse off. And for a big thing—even a very good big thing—the number who are worse off is very unlikely to be zero.
This means that if you want to do big things, you will either have to make some people worse off, or rearrange the gains to make everyone better off.
If there are only a small number of people involved, you might be able to make everyone better off with a careful choice of things to change. But if the group is large, you will probably need some sort of generic value fluid, that can flow between the parties and fill in the holes such as to make everyone a bit better off, instead of some people much better off and some people worse off. Money and social respect both fill this role, assuming that there aren’t other impediments to using them, but a giant barrel of compensatory apricots might also work.
This suggests that whether big changes are made depends on the availability of workable value fluid, along with the propensity of the powerful to make the less powerful worse off without compensation. The availability of workable value fluid might for instance change according to social or technical technology for maintaining it, as well as impediments to using that technology.
For instance, if a large group of people were already headed to restaurant A, but the group would on net prefer restaurant B, they might not make this switch, because someone who prefers B would have to raise the issue, and it would feel a bit too much like conflict (and annoyance of extra negotiation for everyone). However if a couple of the people who prefer B actually own B and can offer drinks on the house to the group—and that is enough for everyone to prefer B, including the B owners—the switch can happen more easily. (I’m really thinking of things like shifts in legislation or giant infrastructure projects, but much more of my own experience is with groups going to restaurants.)
Is this right? Is it a big factor? (Theoretically salient mechanisms can be pretty minor in the real world.)