Breakthrough innovations and welfare: The role of innovators' loss aversion and experience.

Technological refinements appears to be much more frequent than breakthrough innovations. We argue that this could be the result of an optimizing choice when the innovation revenues are exposed to Knightian uncertainty and innovators are loss-averse. The innovator's choice between breakthrough and incremental innovations is analyzed in the context of a neo-Schumpeterian growth model that accounts for the introduction of new goods and related sunk costs. The results show that the welfare generated by breakthrough innovations drops dramatically when agents are uncertainty-averse and/or loss-averse, but rises as innovators' experience increases.