So MVTR is a better predictor of MVTR than air permeability is.
It does all seem a bit obvious. I believe the answer to the question really is purely definitional since I don’t see how your real world test measures anything other than MVTR.
Water molecules are smaller than air molecules, so increasing air permeability must necessarily increase MVTR. And the extent to which increasing air permiability decreases the weight of water on your underlayer should be reflected in the MVTR number. Even the effect of unzipping vents should be explained by its effect on MVTR (it lets air in/out but water vapor too).
I think the confusion may come from the common idea that the water kind of “dissolves” into the air and gets carried away by it, but, from what I understand, this is not the case and water vapor and air should be thought of as just a mix of two types of molecules bouncing around occupying the same space. Your emphasis on vapor pressure implies your understanding of this though, so maybe I am missing something and my line of reasoning is incorrect.