Empirical Studies of IP Protection for Software

By TAP Guest Blogger

Posted on May 3, 2016


This is the final report in the 7-part series from The 20th Annual BCLT/BTLJ Symposium. The reports were written by Kevin Hickey, Microsoft Research Fellow, Berkeley Center for Law & Technology at Berkeley Law School.


A Report from the 20th Annual BCLT/BTLJ Symposium, Part 6


Turning from legal abstractions to practical realities, a number of presenters at the BCLT/BTLJ Symposium tried to better understand the facts on the ground regarding how technology companies and others actually use intellectual property.


Professor John Allison presented years of research that he and his co-authors have conducted on how software patents compare to an “average” patent. (Professor Allison defines software patents as those with “claims covering data processing.”) In a 2007 study with Ronald Mann, Professor Allison found that software patents scored higher than average on a number of measures associated with patent quality. For example, software patents tend to have more claims, more prior art references, and more forward citations, and are patented in a larger number of countries outside the US. In subsequent work, Professor Allison has found that the pattern is true of Internet business method patents, and that these patents are litigated almost 10 times as frequently as a typical patent. While these findings suggest that software and Internet business method patents may not be as low in quality as they are often perceived, these empirical findings do not speak directly to issues of causation. One can argue, for example, that measures like prior art citations are poor proxies for patent quality. Similarly, it may be that patent prosecutors simply have to work harder to obtain a patent in this field because of the various legal obstacles.


Professor Bronwyn Hall surveyed research by economists and others attempting to answer the key question in software patent policy: do such patents actually encourage enough innovation to justify their costs? In theory, the patent system offers social benefits by incenting investment in research and development and encouraging public disclosure of novel ideas (which might be kept as trade secrets otherwise). On the cost side, however, patents raise prices for consumers, impose barriers for cumulative innovation, and fuel costly and often wasteful litigation.


Professor Hall, of course, could not answer this question definitively, but the research that she presented offers some fascinating insights. Studies from the 1990s suggest that firms did not place much value on software patents—market prices for technology firms actually declined in response to the Alappat decision permitting software patents. Evidence from venture capitalists reveals that start-up funding is more difficult to obtain in patent-heavy fields (a.k.a. “thickets”), suggesting that patents impose costs on subsequent innovation. Business method patents, in particular, create heavy litigation costs, being the subject of lawsuits around 30 times as often as other patents—a phenomenon that may be driven by the rent-seeking behavior by patent trolls. At the least, this evidence suggests that software patents are not an unalloyed benefit for software industry.


Professors Stuart Graham and Pamela Samuelson presented the results of 2008 survey of 1332 startup technology firms. They found that patent ownership was much higher among startups with venture capital backing, with about 75% of the businesses in this group holding patents (compared to 25% of non-VC backed startups). When asked about their motivation for patent holding, the most cited reasons were protection from copying and obtaining investment; obtaining direct revenues from licensing IP was cited comparatively rarely. In the field of software and Internet startups, the study also found significant differences in how startup firms view the importance of copyright, patent, and other means of obtaining competitive advantage. Overall, the first mover advantage (getting a product to the market ahead of competitors) was considered the most important, with copyright second in significance. Patent was typically viewed as less important, especially among non-venture capital backed firms. Professor Samuelson added that, with the many important legal and business changes in the field since 2008 (Alice, Oracle, the rise of the cloud, open source, and the Internet of Things, et al.), it may be time for an updated survey of startups to see how their views on IP have changed.


Professor Colleen Chien presented tentative results of her forthcoming study examining the software market through the lens of patent licenses and transfers. Her data set comprised the “material technology agreements” reported by publicly traded companies to the SEC in 2012–14. Professor Chien found that fully 42% of these agreements involved software, and agreements were more numerous in the software industry than biotech or traditional manufacturing. Patents tend to play a mixed role in software agreements, transferring legal liability in a majority of cases, and the technology itself in a substantial minority. Patent rights (key components in 43% of agreements) were more frequently licensed than copyrights (33%) or trade secrets (41%). Many of the biggest agreements were transfers of patent rights from larger, established patent holders (e.g., IBM) to younger companies with fewer patents (e.g., LinkedIn).


Dan Lang of Cisco noted that many of these licenses are more about defensive patent acquisition than the actual transfer of technology. That is, younger businesses are seeking to build up patent portfolios primarily to protect themselves from litigation. Google’s acquisition of Motorola is a well-known example of this phenomenon. Other times, the transaction is really about acquiring source code or other valuable trade secrets—in which case, the transferee obviously also needs any patent rights that go with the code. In general, Mr. Lang sees the level of patent troll litigation as high; this is the environment in which patents are licensed and transferred. Whether Alice will truly change this situation remains to be seen.


Related Resources

John R. Allison – Some Things We Know About Software Patents From Empirical Research

Colleen V. Chien, The Market for Software Innovation Through the Lens of Patent Licenses and Sales, 30 Berkeley Tech. L.J. (forthcoming 2016) (draft) & slides


Read more on the BCLT/BTLJ Software IP Symposium


This report is written by Kevin Hickey, the Microsoft Research Fellow at the Berkeley Center for Law & Technology, UC Berkeley School of Law. Mr. Hickey’s scholarship on intellectual property law and innovation policy is available online at SSRN.