One can believe it test scores correlate with innate intelligence and also think that they are a flawed measurement tool
Sure, but that leaves your argument about the innateness of the characteristic it's measuring as handwaving and special-pleading. You see the same effect with SES in twin studies, for what it's worth.