Does the comparison not work if the smaller city is in Europe and he major hub is one of the ones in the US? Or even two US cities?
I think perhaps using London as an example is overshadowing the point being made. If you aren't aggregating passengers at hubs which you then use larger planes for the common leg of the journey, you don't need as many larger planes.
Direct flights to more destinations from any hub still means a superlinear increase in the number of slot pairs required. Are you suggesting that that wouldn't also be problematic at hubs in North America?
I'm suggesting, as I took sjm-lbm to be originally, that there are more direct flights that are not using the hubs, therefore less need for larger planes to support aggregated passengers. If flights from New Orleans to New York used to go through Atlanta and the Atlanta to New York flight was full of aggregated smaller flights to Atlanta, that would require a larger plane. If more flights are going direct from New Orleans to New York using smaller planes, that lessens the aggregation in Atlanta, and lessens the need for another 747 at that location.
