In the 1930s, Dunlop and Tarshis observed that the correlation between hours worked and the return to working is close to zero. This observation has become a litmus test by which macroeconomic models are judged. Existing real business cycle models fail this test dramatically. Based on this...