In nineteenth century America, ordinary farmers and mechanics wanted to be inventors. Tens of thousands followed the latest developments in periodicals such as “Scientific American” and tens of thousands filed their own patents on mechanical devices ranging from textile spindles to corn shellers. Although few were very successful, many earned rewards selling or licensing their patents to others. The patent system fulfilled its social role by providing critical incentives to startup innovators.
So it is striking that most software developers — the inventors of today’s digital technologies — do not want patents to be granted on software.
This is what large majorities of developers repeatedly tell pollsters. Developers at Silicon Valley companies even protest when their own employers “weaponize” their inventions. Is software somehow different from the mechanical technologies of the past in a way that changes the role of patents? Yes. And because of this difference, software patents do not, on balance, work to the benefit of today’s startup developers.
A flood of litigation
One reason software developers are skeptical about the need for patents is that the software industry grew up without them. The software industry was highly innovative before large numbers of software patents were granted following a 1994 court decision. Software startups were able to make money without patents by being first and best — think Facebook. That is still true today; eighty percent of software startup firms do not obtain patents.
While startups apparently do not benefit from patents on software, other firms do. Indeed, nearly 70,000 patents on software were granted in the US last year, mostly going to large hardware and software firms. Large, established firms acquire tens of thousands of patents in order to battle each other as in the hundreds of lawsuits spawned by the “Smartphone Wars.” Even companies such as Microsoft and Oracle, who grew up with few, if any, patents, now obtain thousands each year. These arsenals of patents may be valuable weapons to large firms, but it is less clear whether they do anything to encourage innovation, the ostensible purpose of the patent system.
Indeed, there is a reason to fear that all of these software patents may actually be discouraging innovation: they have created a flood of litigation. The number of patent lawsuits has quadrupled since 1990 and much of this is due to patents on software, including patents on methods of doing business and mental concepts. A majority of these lawsuits are filed by so-called “patent trolls” — companies who buy up old patents that can arguably be interpreted to cover new technology. The trolls assert these patents, demanding money in exchange for a promise not to sue. Almost all of the patents that the trolls use are software patents; almost all of the companies they sue are either tech companies or customers of tech companies.
And the cost of all this litigation is steep. In 2011, patent trolls in the US cost firms $29 billion and this does not include lost business nor does it include the costs of non-troll litigation. Most of the targets of patent trolls are, in fact, small firms. It is not unusual to find small tech startups in the US dealing with a half dozen or more patent troll assertions. Managers of startups complain of the hours and money they spend dealing with trolls and also about losing business when their customers are threatened.
Unfortunately, there is little that firms can do to avoid becoming a target. It is difficult to determine in advance what patents might “read on” a startup’s technology. There are far too many patents to search, they are often vaguely worded, and the courts are notoriously unpredictable interpreting their boundaries. For example, one inventor developed a kiosk to produce digital music tapes in retail stores and was granted a patent in 1985. But that patent was drafted with vague terms such as “information manufacturing machine.” The patent owner demanded and received money from hundreds of digital e-commerce companies beginning in the late 1990s, even though the original invention was something entirely different. Software patents too often have “fuzzy boundaries,” making it impossible for startups to avoid.
This is what is different about software. Patents are property rights and property rights require well-defined boundaries to work well. This is much easier done when those property rights are based on tangible mechanical devices or a chemical compositions than on abstract mental concepts.
Bad fences make bad neighbors
And so software patents are four times more likely to be in a lawsuit than are chemical patents; patents on methods of doing business are thirteen times more likely. Bad fences make bad neighbors, allowing patent trolls to exploit poorly defined boundaries.
Hence software startups face a substantial and unavoidable risk of litigation. Faced with this risk, they will not stop innovating. But they will invest less and they might not make some risky innovations at all. The risk of litigation acts as a tax on innovation. Given that these startups receive little benefit from patents, software developers have good reason to feel differently about patents than their nineteenth century counterparts. Instead of encouraging startups to innovate, software patents today discourage them. Perhaps legal doctrines can be tweaked to increase the clarity of legal boundaries for software patents. Perhaps. But until judges recognize how software is different, the problem will continue.