Optimal production of innovations under uncertainty

The paper reports results on a risk-neutral firm's research incentives. When unrelated to the firm's own stake in the program, the risks encourage or discourage risky research spending, depending on the properties of the research technology available. A non-decreasing time path of information builds the idea of an asymmetric probability distribution of the state of knowledge into the model. It follows that the required return on risky investments may actually fall short of the safe return. Since it is the upside risk that dominates, increased controllable risks will increase incentives for risky innovative activity. It is proved, but only in a more restricted framework (with differentiable processes), that the expectational effects involved will strengthen the positive relationship between controllable risks and the expected return.I am greatly indebted to three anonymous referees for helpful suggestions and to the Yrjö Jahnsson Foundation for financial support.