Non-Obviousness and Complementary Innovations

I am grateful to the publisher, Elsevier, for letting the manuscript being archived in this Open Access repository.International audienceThis paper explores which patent policy should be applied to complementary innovations that are aggregated into broader technologies. I compare a setting in which complementary innovations must be bundled prior to patenting, with a second setting in which they can be patented separately. The first setting can improve static efficiency by avoiding the costs resulting from the scattering of complementary patents. But it also limits the disclosure of small innovations, which may lead to inefficient R&D cost duplications. A model capturing these effects shows that patenting complementary innovations separately is not efficient when innovations can be developed rapidly. This result justifies the enforcement of a severe “inventive step” or “non-obviousness” requirement in sectors where complementary innovations are frequent