Revisiting the Private Value of Scientific Inventions

Estimating the private value of patents is important, yet challenging. By developing a method that uses stock market returns to produce a distribution of patent values (and not just an estimate of the mean of that distribution), Kogan, Papanikolaou, Seru, and Stoffman (2017) (KPSS) opened venues for new research. Researchers have used these estimates to compare average values of different types of patents. In this paper, we argue that KPSS values should not be used in their current form to compare mean values of different groups of patents, and show this to be the case in the context of research on the private returns to scientific patents. We extend the original KPSS method to allow for patents to be drawn from two distinct value distributions. Using this approach, we find that scientific patents have a higher mean value compared to non-scientific patents.