This dissertation explores the costs and benefits of inducing innovation at the level of markets, individuals and knowledge. First, I examine the extent to which the National Institutes of Health can direct its resources and generate production in specific areas of science. I find that it can, and that science funded through these directed efforts is significantly more productive than average. Second, I identify how willing individual scientists are to adjust the trajectories of their research in exchange for additional resources - the elasticity of direction. Estimated magnitudes suggest that the directional adjustment costs of biomedical science are large enough to warrant policy attention. Finally, in joint work with Mark Pauly, we explore the growing costs of R&D in the pharmaceutical industry in order to identify how much might be efficient cost growth in response to a larger market. Almost all of the growth in R&D spending can be attributed to demand, with no evidence that marginal investments have become less productive over the past thirty years.