Portfolio Choice in Research and Development

We analyze the effects of a "winner-take-all" patent mechanism on the riskiness of the research strategies chosen by competing firms, as well as on the firms' incentives to duplicate research projects. Nash equilibrium choices are compared with the social optimum in a one-shot, simultaneous-move game in which competitors choose the riskiness or correlation of their research performances. We show that neither society nor firms have any preference for correlation per se, while the divergence between social and privately optimal levels of risk depends on skewness characteristics of the probability distribution over the discovery dates and on levels of risk aversion.