Skip to main content

How to Kill a Patent Troll

Researchers tackle whether “non-practicing entities” are just litigious trolls.

By Nathan Collins


Sitting in a small, fifth-floor office on the edge of San Francisco’s Civic Center neighborhood, it’s just as easy to see the dark side of technology—the arrogant belief in the inherent good of the latest smartphone application—as it is to see the shiny, happy optimism of innovation. But there is another monster lurking in the shadows, a drain on innovation that we are only just starting to comprehend: the patent troll. And according to a perspective published in Science last week, it needs to be stopped.

For the uninitiated, patent trolls are businesses that exist for the sole purpose of suing other companies for violations of their intellectual property. Formally known as non-practicing entities (NPEs), they don’t actually manufacture or develop anything, nor do they even come up with their own ideas. Instead, they acquire others’ intellectual property, then sue companies for supposed violations.

On paper, NPEs might not deserve the “patent troll” epithet. They could, in theory, exist to protect the rights of the backyard inventor from the Goliaths of the tech world. “A conflicting hypothesis suggests that NPEs are patent trolls,” write Lauren Cohen, Umit Gurun, and Scott Kominers, able to extract money from more productive companies who would rather settle than deal with the headache of litigation, regardless of who’s right and who’s wrong.

NPEs go after companies with the biggest wallets and the fewest available minutes.

Anecdotally, NPEs are trolls. But Cohen, Gurun, and Kominers wanted some hard proof. For that, they turned to data from RPX Corporation, which maintains a database on NPE litigation going back to 1977. (RPX also offers its clients a novel and slightly odd solution to patent trolling: It buys patents from NPEs before they start suing others for licensing fees. RPX asserts they are not themselves patent trolls.)

Both the RPX data and other sources make it clear that NPEs are predominantly trolls, mainly because of who NPEs go after: cash-rich tech companies. Cohen, Gurun, and Kominers calculate that the likelihood of getting sued by an NPE is roughly 16 percent among companies with the most cash, roughly double the baseline rate. By comparison, the likelihood of getting sued by a practicing entity—that is, a company that actually worked to create its patents—is less than five percent. NPEs are also more likely to sue firms with small legal teams and those dealing with other lawsuits. In other words, they go after companies with the biggest wallets and the fewest available minutes.

The consequence is a decrease in research and development investment of around 25 percent, the researchers argue in a related working paper, though they caution that it’s hard to identify exactly why companies reduce their R&D budgets following intellectual property litigation.

Part of the solution to that problem may lie in reducing the cost of challenging NPEs’ dubious patents. “Additionally, mechanisms should be adopted to weed out low-quality patent lawsuits,” the team writes. “A system of ‘advance screening’ could require that patent litigation be preceded by a brief court appearance and/or patent review. The review would provide preliminary evaluation of whether the plaintiff’s infringement claims are reasonable and of whether the asserted patents are of high quality.”