Innovation is central to modern economies. Governments recognize innovation’s importance for growth and job creation. Firms view innovation as a key source of competitive advantage. However, despite innovation’s broad importance, relatively little is known about how to foster innovation. Does innovation require long term commitment to a single idea or rapid “pivoting” between multiple ideas? Should government strengthen or weaken property rights to improve an industry’s rate of innovation? How important is team composition to a group’s ability to innovate? 

Empirical work addressing these kinds of questions have been stymied by the nature of innovation itself. Innovations are unique. They can differ in aspects such as the benefits they generate for users, resources required to bring to market and products they could improve. This poses a challenge to the empirical study of innovation by calling into question whether any dataset of innovations allows for statistical analysis. Causal statements of why some environments led to increased levels of innovation would be confounded by variation in the type of innovation generated in each environment. Hence studies of innovation have been limited to datasets such as patents which allow researchers to control for some of this variation. Although studies of patent databases have produced valuable insights on innovation, decreasing returns to future work call for new sources of data on innovation. 

To that end Ankur uses the television industry to expand our understanding of innovation. Each television show can be viewed as an innovation since shows are unique, each bringing a new product to market. The production of television shows also has a set rich of characteristics useful to studying innovation. Sometimes the first episode of a show is produced before the full show is ordered, allowing studies of how commitment affects innovations. The intellectual property of shows can be held by the distributor or by 3rd parties, allowing studies of how property rights affect innovation. The teams behind each show can be tracked across their entire careers, allowing studies of how team composition affects innovation. Over the industry’s history, entrants, technological change and regulators have generated shocks to the industry which could enable causal empirical research.