Empirical work and evidence genuinely matter and can change minds (as the debt-relief data changed Henry's), but the profession's obsession with multivariate regression and instrumental variables to tease out causation has been an intellectual dead-end even while being productive for careers.