The recent NHL Stanley Cup Finals made me ponder the question of league geographic choices. Detroit (Western Conference) and Pittsburgh (Eastern Conference) are only a little over 200 miles apart. Most observers thought the Red Wings-Penguins to be a dream match for hockey fan interest, and it may have been. However, there are instances where such close geographic proximity has diminished broader national or regional interest.
Sports league differ considerably with respect to geographic division and, therefore, how championship games or series are decided. The NBA and NHL have two conferences roughly dispersed along East-West lines. Of course, Detroit's residence in the West muddles this distinction. In contrast, the NFL's conferences and MLB's leagues are dispersed geographically at the divisional level only so that two teams from analogous regions can meet in the championship (Giants-Patriots, Eagles-Patriots, Giants-A's, Yankees-Mets). Again, there are geographic inconsistencies such as Dallas in the East Division of the NFC.
There may be sound reasons for geographic breakdowns differing between sports. The overall level of fan interest comes to mind-- as the alpha dog of sports, the NFL may be able to sustain national interest regardless of the proximity or distance of the team's playing more so than sports with lower levels of support.
League politics, and not merely economic influences, likely plays a role. Geographic realignments usually require both broad agreement of owners as well as agreement by specifically affected parties. This presents a high "transaction cost" of negotiating new alignments, so that old alignment structures probably stay in place well beyond their economic value. For example, MLB once held top status among American sports with the World Series garnering widespread attention even if two teams from New York played. Now, even bad Monday night football games can outdraw the World Series. With less overall fan interest, close geographic matchups may hurt the league.
Perhaps some SE readers or contributors know of studies on this topic. If so, I would be interested.