Who Wins and Who Loses When IP Protection Gets Weaker?

That protection flows from the Patent and Copyright Clause of the Constitution, which grants Congress the power “to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.” Congress first codified patent protection in the Patent Act of 1790. Section 101 of the most recent Patent Act lists the subject matter that can be patented: “any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof.”

A large body of Supreme Court jurisprudence has interpreted the Patent Act over the years, carving out exceptions to the law and further clarifying the broad language in the statute. In the 2014 case of Alice Corporation v. CLS Bank International, the Court dropped something of a bombshell on the patent law landscape. It ruled that patent claims regarding computer-implemented inventions—including systems, machines, processes, and items of manufacture—are not eligible for patent protection if they are mere abstract ideas that do not transform into something novel. The Alice decision immediately threatened firms in several patent-heavy industries, such as data processing, software, and measuring or testing in microbiology and enzymology, calling into question their existing patents and the firms’ ability to secure new ones.

While the Alice decision has been the subject of much study by legal scholars, the ripple effects of the decision on firms and innovation have been difficult to quantify. That’s the challenge Tuck professor Gordon Phillips and colleagues take on in “Intellectual Property Protection Lost and Competition: An Examination Using Machine Learning,” a new working paper that uses a novel machine-learning methodology to analyze a trove of patent data. The paper studies how the Alice decision may impact firms in two main areas: by potentially revoking or denying patents that existed or were applied for before the decision, and by impairing future efforts at patenting, sales growth, and profitability.

To determine which firms may be vulnerable to the new criteria from the Alice decision, Phillips and his co-authors examined the text of patents and patent applications between 1994 and 2014. That yielded 3.8 million patents, so they narrowed their analysis to only those patents that have the same Cooperative Patent Classification as the patent applications that were rejected after 2014 according to the Alice criteria. This narrowed their sample size to 642,678. Then they used “a deep-learning-based language model called BERT to predict the likelihood each of the pre-Alice granted patents in the sample may be invalidated by the Alice decision,” the authors write. The BERT model was released by Google in 2019 and stands out in its ability to process words in relation to all the other words around them, thereby achieving an understanding of the context of the words.

Once they identified the firms vulnerable to the Alice decision, the co-authors examined the impact of the lost intellectual property protection on firm innovation, performance, competition, and mergers and acquisitions. They found that post-Alice patenting significantly decreases for both large and small firms, and a significant increase in R&D for small firms. “These results are consistent with small firms’ attempting to replenish their innovative portfolio and to rebuild product differentiation after losing IP protection on their existing patents,” the authors write.

When it comes to post-Alice changes in sales growth, profitability, and firm value, they found an asymmetric impact: large firms win and small firms lose. Exposed large firms gain in sales and market valuation, while small firms see a decrease in operating margins and market valuation.

What might account for this asymmetric impact? The authors show that it is explained by changes in competition for small firms—including increased venture-capital-financed entry into their market, and lost product differentiation—and limited legal options to replace losses in IP protection. Large impacted firms were also less likely to buy smaller firms.  They documented these small firms resorting to non-compete and non-disclosure agreements as substitutes for patent protection.

Given that the Alice decision particularly injures small firms, the paper highlights the “benefit for small firms of removing the ambiguity of Alice or even repealing Alice as a recent bipartisan bill attempted to do,” they conclude. They also note that there are some benefits of Alice and the loss of IP protection—there is increased competition and less lawsuits targeting large firms by non-producing entities which some commentaries have called patent trolls.