if you read the actual research provided by the OP it shows way more than that. The measures did not only crowd out hate speech, the study followed individual users and noted that they moderated and changed their behavior once they moved to other communities after the more extremist ones were banned, and they individually changed adapted.

That implies that the there's a positive effect even on the individual level, not just a silencing effect. Which seems completely in line with say, busting a cult in real life.