Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This article studies how a firm's reputation for rewarding innovative employees affects innovation and startup creation. In any Pareto-efficient equilibrium of the repeated game, low-value innovations are developed in-house, while high-value innovations are developed in startups. When distributions of ideas are ordered by simple cases of first- or second-order stochastic dominance, the firm has a preference for an extreme distribution. The article also characterises the optimal relational contract and workers' incentives to invest in innovation. The model's predictions are consistent with a broad set of observed regularities regarding the creation of employee startups.