In a two-stage framework, a manager delegates invention and makes innovation decisions after observing the quality of the invention. The optimal contract for delegated R&D is shown to take the form of an option. The analysis extends the principal-agent model to allow the principal to implement the invention ex post. The analysis further extends the principal-agent model to consider experimental design with simultaneous sampling and sequential sampling. The discussion introduces a general "implementation principle" that simplifies the analysis of moral hazard models and eliminates the need for monotonicity assumptions either on the contract or on the principal's benefit function. The model extends to various important statistical decisions: when the agent designs the experiment by choosing the distribution from which to sample, when the principal imperfectly observes the outcome of R&D,and when the number of samples is random.
|Original language||English (US)|
|Number of pages||37|
|State||Published - Jul 28 2014|