Working paper (with Josh Feng) available upon request.
Using patent prosecution data, we shed light on the patent acquisition behavior of non-practicing entities (NPEs) and identify policy levers within the United State Patent Office (USPTO) that can help limit their activities. We nd that NPEs, unlike regular rms, purchase and assert patents that were granted by a specic set of examiners, who tend to allow obvious patents with vague claims. Our methodology is based on the random assignment of patent applications to examiners and provides a novel way of inferring the nature of a patent from prosecution data. A simple cost-benefit calibration based on the fees from NPE litigation suggests that policies increasing the quality of the patent examination process at the USPTO could have very large social returns.