We estimate the relative contribution of mobile scientists who leave academia for the private sector on the subsequent innovative performance of the firms they join. We use data on the population of Danish firms and their R&D workers for the period 1999-2004 and measure innovation performance by the (value-adjusted) number of patent applications at the European Patent Office. We compare the efficacy of mobile former university scientists to the effects of mobile workers hired from other firms as well as immobile workers on the innovation performance of their employer. Our main result is that mobile university scientists contribute substantially more to innovation than R&D workers hired from other firms who, in turn, contribute slightly less to industrial innovation than recent university graduates. By contrast, immobile workers add little to the innovative activity of their employer. We also find that the contribution of mobile R&D workers to innovation depreciates fairly rapidly. These findings provide us with three main managerial implications: Firstly, hiring scientists from universities is a way of boosting a firm's innovative activity. Secondly, because hires from academia receive lower wages on average than hires from private sector firms, this implies that hiring R&D workers from academia may be a cost-effective way of improving innovation performance. Thirdly, firms need to take measures in order to further public-private researcher interaction to prevent the depreciation of the knowledge stock of their employees.