The doubling population of Kenya suddenly seems important to you. Your accusation that I neglected the population growth is plain wrong. In the very second post, before giving any summary quantity and quality data, I gave percentages of the world population for both 1990 and present, reflects the increasing population of East Africa and North Africa, with respect to global population:
As much as Kenyan population has grown, "5 continents" is still much larger than all of East Africa by a factor of ~16, and larger than all of North Africa, by a factor of ~29:
The United States outnumbers the Kenyans by about 6.5x. Do you think there are 90 Americans today, who if given unlimited access to Kenyan drug stores, would outperform Coe of 1985? No American in history has done it, unless you count Lagat. This is a part of what you are calling the obvious higly likely scenario that I reject.
Regarding your second point, why do you say "Rekrunner's 'talent has moved to the roads' maxim doesn't really cut it, does it?" Because of Kenyan population doubling? Why do you reject "Money dried up on the track" as a solid reason?
When I looked at the marathon, I counted 100 "quality" performances (sub 2:07:35) (62 performers) before Sep. 2007 (over 22 years), when Geb set his first world record, and 447 "quality" performances (202 performers) after (just over 10 years).
When I counted sub-2:07:11 (my 1990 benchmark) performers, this was from 183 East Africans, 6 North Africans, and 12 from the rest of the world.
Carlos Lopes ran 2:07:12 and Steve Jones ran 2:07:13 in 1985, so 2:07:11 should be quite feasible for all populations over the next 33 years.
This looks like a strong indication to me that a lot of East African talent has moved to the roads after 2007.