On the forty-second page of “The Free-Market Innovation Machine: Analyzing the Growth Miracle of Capitalism” Princeton & NYU economist William J. Baumol wrote (emphasis added):
be explainable by the Ricardian rent model, with rewards to the different supermarginal innovations corresponding to differences in the market values of their degree of superiority over the marginal innovation. Any economic profits that accrue to the incumbents in the industry will, then, be protected from entry so long as they do not exceed the minimum amount needed to attract new firms into the field, given the risks entailed in any required sunk investment. For this reason, the zero profit conclusion is considerably weakened.
Still, casual observation suggests that there is a good deal of truth to that story, as we have seen. The most extreme cases of routine and predictable innovation—the annual introduction of new automobile models or new spring clothing fashions—do not seem to promise vast economic profits to the participants as a group. This seems even more so for the continuing innovation wars that characterize the computer industry. At the very least, such impressionistic evidence serves as a warning against easy acceptance of the early Schumpeterian view that investment in innovation is a reliable source of economic profits.