Like much else in Keynesian economics, this hint of empirical evidence is erroneous as a general theory.