Moral hazard in innovation: the relationship between risk aversion and performance pay

Standard principal–agent models study the situation in which there is moral hazard within a given production method. The classic prediction of these models is that incentives tend to decrease with risk and risk averseness of the agent. The prediction is inconsistent with empirical findings, which reveal that both negative and positive relationships between risk and the power of incentives co-exist in the real-world. Our paper develops a tractable model to illustrate that introducing innovation effort could well explain the puzzle of the relationships between the power of incentives, risk and risk-averseness of the agent. In particular, we assume that the agent can exert two unobservable efforts: a production effort and an innovation effort. While the production effort can increase output within a particular production method, the innovation effort can generate an innovative but riskier production method. We show that, when inducing innovation is unnecessary, the standard result that incentives tend to decrease with risk and risk averseness holds. However, when inducing innovation becomes necessary, incentives may increase with risk and risk averseness. The insights in this paper are especially important today, when innovative but riskier technologies upgrade very quickly.