The Allocation of Scientific Talent

I explore the efficiency properties of a decentralized labor market for scientists. I use a model where firms produce science by building labs and hiring researchers in a competitive market. Firms may invest in science to produce new scientific knowledge or to increase their absorptive capacity: the ability to use scientific knowledge produced outside of the firm. In both cases firms underinvest in labs. More interestingly, when firms' investment in science is motivated by absorptive capacity, researchers and labs may be substitutes in the revenue function, even though they are complements in the research production function. This generates a novel form of inefficiency: for any given distribution of labs, the allocation of researchers to firms is non optimal. Subsidies to the investment in labs cannot restore the first best. I show that the existence of scientists' reputation concerns, by preventing the free transfer of surplus between firms and researchers, may affect the allocation of scientists to labs and increase total welfare.