I agree with that predictive ability is important but I have seen no evidence that marginal economics has predictive ability. You yourself even say it "explains and models" which is not the same as predicts.