Indeed, the myth of equality across racial and cultural lines must be protected at any cost.
and trying to tease out its meaning in the context at hand.
The context at hand appears to be in making decisions about people's lives. I don't see how to read it in any other way than "people should discriminate on the basis of race and culture"; and given reinhardt's previous comment against positive discrimination, he can only mean negative discrimination. Statistical equality does not seem relevant at all; I don't know why you brought it up. I'm suspicious that you are dissembling.
I interpreted "equality" in reinhardt's statement to mean statistical equality. Statistical equality is an empirical claim - if people wrongly believe it to be true, it would be a "myth".
No, the implication is that P(inferior | race X) != P(inferior | race Y).
...we do hold to equality across racial and cultural lines;
There are two types of equality under discussion:
a) Statistical equality - P(Q > cutoff | race X) == cdf(Q > cutoff| race Y), where Q is some quality measure which is not a function of race.
b) Equality of treatment - using the same cutoff value for race X and race Y.
Which type of equality are you referring to?