The Bigger the Group, the Less Likely Pareto Improvements?

All classes were canceled at the University of Texas one recent Tuesday because of a shooting at the main library. I already had a haircut appointment scheduled for 4 p.m. and suddenly had a lot of free time before then. I called the hairdresser and asked to come by at 2. She was free then. This gave me a preferable work schedule; and, as I had been her last appointment, she got off work early.
A clear Pareto improvement – if the relevant set of people is my hairdresser and me. If the set is expanded more broadly, students were probably better off; the administrators and police who had to deal with the crisis surely were not. In general, if you expand the set of concerned parties to any transaction even a bit broadly, I wonder how many cases there are when Pareto improvements are possible. We teach this concept to our freshman, but increasingly its usefulness escapes me.