> And other approaches are significantly more useful? They yield massively more predictive power? They are more tested and less theoretic?

Steve Keen's own theories (well, he builds on Minsky, mostly) can predict how great moderation suddenly bursted into debt deflation, explain high empirical correlation between debt acceleration and unemployment, show that quantitative easing works better than austerity in debt deflation (and predicts that helicopter money would work even better, although nobody tested that) - this was empirically proven comparing strategies of Japan, U.S. and EU in crisis. He can also explain other things that leave most macroeconomists puzzled.