Abstract: | I consider a model in which firms produce new knowledge by building laboratories and hiring researchers in a competitive market. I show that, for a given distribution of laboratories, the allocation of researchers to firms can be efficient or inefficient, depending on how fast the firms’ marginal return on the knowledge produced decreases with the amount of knowledge produced. I then argue that the allocation of researchers to laboratories is likely to be inefficient if firms invest in research and development primarily to increase their absorptive capacity (i.e., their ability to use the stock of publicly available knowledge). When the distribution of laboratories is endogenous, a second source of inefficiency arises: firms’ underinvestment in laboratories. Policies subsidizing investment in laboratories are ineffective at restoring the first best, unless policies aimed at reallocating researchers to firms are also put in place. |