I'm saying that through self-selection, lack of alumni, whatever, T14 is for national portability. The few after that (Vandy, UT, UCLA) are for dominant regionals. I'd argue that it should be the T12 for national, and include Cornell and GULC in the dominant regional group.
I generally agree with your premise, that 13-17 don't have as much difference as they're given credit for. I definitely disagree with your assessment that they're strong regional schools, though. A strong regional school is Minnesota, or Alabama, or BU. I was totally with you in your earlier point, that UCLA, Texas, and Vandy suffer the most from self-selection, in that Texas people want to go to Texas and then want to stay in Texas, but it's impossible to definitively determine to what extent those schools would place otherwise, because they don't! If Vandy was in Boston, would it be ranked higher? Probably, but it's an impossible argument to make, because it's in the South, and it attracts southern applicants, at least a percentage of whom will want to remain in the south, but nothing about Vanderbilt is even slightly regional - saying Vanderbilt places primarily in the south is like saying Cornell places primarily in upstate New york - not accurate. And calling Georgetown regional because it places primarily in DC is equally as dubious - by that criteria, Columbia and nyu should be qualified as strong regional schools since the majority of their graduates stay in ny (because why else would one endure the astronomical COL for 3 years if they're just going to leave?)
Basically, I'm arguing both for and against your point - I agree there's negligible difference between 12-17, but not that they're all "strong regional schools"; they are all schools with very national reputations; I'd wager the only differences between those schools and higher ranked schools is the placement for the students below the median, who have murkier job prospects the further down the list.