In this answer I suggested that "the majority" of NPCs would be 0-level characters, and was told by our friendly moderator @BrianBallson-Stanton in a comment (since deleted) that D&D 4 didn't have (or wasn't supposed to have) level gaps.
In AD&D (2nd Edition) campaigns that I ran and played in with my friends, in the absence of any written canon guidelines, we made assumptions about the relative levels of the NPC population of the worlds that we called the "Fifty-Percent Rule"; that the majority (i.e. 50%) would be level 0 characters, and of the other 50%, 50% of them would be Level 1, and so on, so that:
Level 0 : 50% Level 1 : 25% Level 2 : 12.5% Level 3 : 6.25% Level 4 : 3.125% Level 5 : 1.5625% Level 6 : 0.78125% Level 7 : 0.390625% Level 8 : 0.1953125% Level 9 : 0.09765625% Level 10: 0.048828125% Level 11: 0.0244140625% Level 12: 0.01220703125% Level 13: 0.006103515625% Level 14: 0.0030517578125% Level 15: 0.00152587890625% Level 16: 0.000762939453125% Level 17: 0.0003814697265625% Level 18: 0.00019073486328125% Level 19: 0.000095367431640625% Level 20: 0.0000476837158203125% and so on...
Referring to the answer I gave, for a D&D 4e caster to be able to cast Raise Dead, a Level 8 spell, they would need to be 8th level at minimum, i.e. in the top 0.390625% of the population. In a city with a population of 10,000 (Not unreasonable for a D&D-type world), this would be 39 individuals of any character type of Level 8+. Given that most of them (say, 90%) would be unable to cast the spell in question, there may be at most three or four (3.9) who could actually Raise Dead.
My question is:
Is this a reasonable breakdown of the population distribution of experience levels in a D&D campaign (any version)? Is there some canon reference in any version of D&D that states an actual level-population distribution?