I didn't say that. I said that the 13.4%/23.4% data is reason for deeper inquiry.
In a civil context - like the mortgage lending class action case I was talking about - that alone would be sufficient for a prima facie case of discrimination. It would shift the burden to the defendant to prove a business necessity - such as incomes, or loan-to-value ratios, or debt-to-income ratios - are different between white and Black borrowers. BTW, any of those business necessity factors could be influenced by systemic racism . . . the question then would be whether the defendant has direct responsibility for the statistical discrepancies present in those factors.
This approach is much more difficult in a law enforcement context . . . but it needs to be done, if in a modified fashion. In a law enforcement context, the police (rightly or wrongly) are proxies for the society at large. That's why the individual officers can be not charged, but the government and police agree to make changes such as no more no-knock warrants. But the overall accountability remains . . . .