It is argued that the linearity of the R&D sector's technology in Paul
Romer's model of endogenous technological progress is very restrictiv
e. An externality, stemming from the simultaneity and overlap in resea
rchers' activities, is modeled into the R&D sector. The optimality ana
lysis of this model leads to a sceptical view of the case for governme
nt subsidies for investments in R&D. The quality of human capital, a s
ubject discussed rarely if ever in theories of endogenous technologica
l progress, is shown to affect the model economy's long-run growth rat
e.