If understand you correctly, your answer to my question would be "never", that is, you would always attribute some blame to historical causes. Okay.
I am left with more questions, however. To paraphrase your final paragraph, you expect that efforts to undo the effects of past racism—those effects which we collectively call systemic racism(?)—is necessary to snip that past racism from the causal chain to present ills. But I'm left wondering if this language of systemic racism is even particularly useful in describing the situation.
That is, it seems the manifestation of this framing is to address these downstream effects (poverty, etc.), none of which are inherently racial, but affect educational outcomes. But it seems to me that framing the problem nonracially and focusing solely on the proximal causes of educational issues has the same (or better!) manifestations as the racial framing.
In short, I feel the systemic racism framing is unproductive, because in a prudent implementation it merely adds discussion of distant causes, while identifying the same social issues to address. In an imprudent implementation, it would not only cloud the field with historical discussion, but distract from important proximal issues which don't fit the historical frame, while at the same time alienating people who feel excluded or infantilized or condescended upon based on their immutable characteristics, which is scarcely outweighed by a possible ethnic rallying effect which could boost participation.
I think I need to provide a concrete hypothetical to tidy up. Consider a cohort of struggling students in Virginia, say, old coal town. The sociologist correctly identifies historical racism as a factor in some of the students' issues. So they... what? Acknowledge it? What for? They begin their real work addressing (somehow, idk) the homework environments kids have, their encouragement to succeed, the parents' support, school supplies, whatever. And race comes into the calculations exactly... never. I imagine it would be very disturbing if it did. "We're gonna help the black kids first because Jim Crow happened and that means they need it more." Well... maybe! Why make the approximation? Just focus on the proximal causes and get a precise prescription, no need for rounding.