Abstract: How do non-practicing entities impact innovation and technological progress? The question has enormous importance to industrial policy, with virtually no direct evidence to inform it. This paper provides new evidence on the subject, both theoretically and empirically. In doing so we inform the debate that has portrayed NPEs alternatively as benign middlemen that help to reallocate IP to where it is most productive or stick-up artists that exploit the patent system to extract rents, thereby hurting innovation. We make use of unprecedented access to NPE-derived patent and financial data as well as a novel model that guides our data analysis. We find that NPEs target patents coming from small firms and those that are more litigation-prone, as well as ones that are not core to a company's business. When NPEs license patents, those that generate higher fees are closer to the licensee's business or more likely to be litigated. We also find that downstream innovation drops in fields where patents have been acquired by an NPE. This evidence provides some support for both views of NPEs and suggests that a more nuanced perspective on NPEs as well as additional empirical work is necessary before informed policy decisions can be made.