This paper conducts various tests of four hypotheses to account for pr
ocyclical multifactor productivity. The hypotheses include imperfect c
ompetition, increasing returns to scale, labor hoarding, and thick mar
ket externalities. The test results provide strong evidence for imperf
ect competition and increasing returns, and weak evidence for labor ho
arding and thick market externalities. However, the markup and returns
-to-scale parameter estimates are much smaller than those noted in pre
vious research. This difference results from the use here of gross out
put and intermediate inputs as opposed to the use in previous research
of value added and primary inputs.