ABSTRACT
Roy Harrod's point of departure was the short-run theory of income determination, originally developed by Keynes as a means of analyzing the causes of mass unemployment in the 1930's. Since investment demand depends on the growth of income, Harrod's model appears to be an example of the Multiplier-Accelerator interaction–but with the special feature that the Accelerator is unlagged. The insertion of a lag into Harrod's investment function, however, changes the model almost beyond recognition. It is hard to believe that upward trends in per capita output are to be attributed entirely to improvements in technology and not at all to increases in the amount of capital per worker–yet this is what is implied by the assumption of fixed factor proportions. On the other hand, to abanduced don that assumption as too restrictive is to change the Harrod theory out of recognition and to take the first step in the so-called "neoclassical" approach to the problem of growth.
