This paper focuses on the issue of risk in the R&D and innovation processes
. It examines the degree of risk suggested by the four main streams of the
literature. While risk has not been a major focus of the empirical literatu
re, the results suggest that the vast majority of inventions are worth rela
tively little, while a small proportion are extremely valuable. The second
part of the paper argues that the net present value rule may not be appropr
iate in the presence of high risk, and the R&D decision may require an opti
mal stopping rule, which gives very different results. Adopting an optimal
stopping rule in the face of a normal distribution of annual gross returns,
the lifetime distribution of returns are similar to those observed in the
real world. In addition, it is possible to use the results to replicate the
observed behaviour of patent renewals.