Rarely a week goes by these days without the publication of one academic study or another that addresses the “problem” of software patents in the US, often in tandem with a dissection of the cruelties inflicted by “patent trolls”. The conclusions of these studies are almost always variations on the same theme: “Something has to be done because the current regime is stifling economic growth, as companies are afraid to innovate because they may be sued for patent infringement.” These studies are always very precise about the specific problem within the system that they address, but they never seem to provide any evidence that innovation is, in fact, being affected by patents. That just seems to be taken as an article of faith.

However, as I have argued many times on here before, that simply is not good enough: no-one doubts that there are uncertainties around software patents, neither does anyone serious deny there are bottom feeding trolls that do pray on non-IP-savvy SMEs in the hope of securing licensing fees covering poor quality patents. Instead the issues are these: How much uncertainty is there? How many bottom-feeding trolls are there? How many companies are affected by them? What percentage does this number represent in terms of the overall number of companies creating and using software? Where is the proof that there would have been more innovative activity if these bottom-feeders had not existed? How damaging are these trolls’ activities? Would legislating against them have the potential to hurt more companies than it would help?

As far as I am aware, no serious attempt has been made to address any of these questions. But given the potential damage legislation could cause to a wide community of patent owners (current and future), it seems to me that both legislators and courts in the US need answers before they begin to contemplate action. In short, what the current software patent and troll debate lacks in the US is any kind of context. And that is a problem, because without context people make decisions based on hunch and anecdote – neither of which is remotely reliable.

This morning I was reading a paper called Scaling the Patent System, published last year by Christina Mulligan and Timothy Lee. It argues that because so many software patents are granted by the USPTO each year, it is impossible for anyone to keep up with what they cover. This creates uncertainty and helps to encourage litigation, which in turn creates disincentives for individuals and companies to innovate. It’s an interesting theory, but from what I can see no practical proof that it actually works in the real world is offered: it is a theory and no more. That said, what the authors do quite correctly state is this:        

The number of firms producing patentable software is massive — much larger than the number of firms in the software industry as it is conventionally defined. Almost every medium and large American firm has an IT department that performs backups, runs file and mail servers, runs the firm's website, and so forth. IT professionals routinely create software to automate such tasks, and this software is potentially patent-eligible. Many firms also develop custom software to automate common business processes, and some of it is quite complex. Hence, most medium and large American firms (as well as many non-profits, universities, and other organizations) are in the software industry as far as patent law is concerned.

They then go on to estimate how many organisations may produce patentable software: “There are roughly nf = 634,000 firms in the United States with 20 or more employees. While not all of these firms produce software, many of the 1.7 million firms with 5 to 19 employees do, so we’ll estimate the number of firms that create software at 600,000.”

Now think about that for a moment. Elsewhere the authors tell us that the USPTO issues around 40,000 software patents each year. If we apply a one patent per entity rule (which is ridiculous, I know, but stick with me), what the authors are saying is that less than 10% of companies that could seek software patents in the US actually receive them. In reality, of course, the percentage is even lower because individual entities receive multiple software patents annually. Thus by the authors’ own figures the vast majority of software-producing companies in the US do not engage with the patent system.

But what of litigation, I hear you ask. Well, we know that PWC estimates that between 1995 and 2011 just under 6% of filed patent suits related to software. If we say there were around 50,000 patent suits filed in the US between 1995 and 2011 (which is probably a slight overestimation), that means around 3,000 software cases; or an average of just under 300 a year. Of course, some of these may have had multiple defendants; while it is also true to say that for every case that ends up being filed many more are settled (though the corollary of that is that the same names appear as defendants in multiple cases).

Whichever way you look at it, therefore, not only are we talking about a small minority of all companies that create and use software in the US seeking patent protection for the fruits of their labour, but also a tiny percentage being involved as defendants in patent disputes. In other words, if we accept the Mulligan and Lee 600,000 estimate, we also have to accept that the chances of a company being accused of patent infringement if it creates and uses software in the US, let alone of it ending up in court, are very low indeed.

Thus, if fear of patent litigation really is harming innovative activity relating to software (though there is no evidence of this), is it not more likely that it is perceptions of the problem which cause doubts rather than actual experience? After all, you cannot blame people for worrying if they constantly read in various media outlets that software patent litigation is out of control and trolls are running amok, even if in reality neither is actually the case. In other words, what may well be causing this unproven software innovation deficit in the US (which may not actually exist) is not software patents or the actions of trolls/NPEs/PAEs, but the huge media coverage given to a series of academic studies that, without providing any practical evidence, claim that there is a software innovation deficit in the US because of software patents and trolls/NPEs/PAEs.

Funnily enough, this narrative fits in very nicely with the vociferous complaints of a corporate elite which frequently end up as defendants in cases relating to software patent infringement brought by other operating companies and trolls/NPEs/PAEs. Someone more cynical than me might conclude that it is in their interests to ensure the seeds of uncertainty and fear are sown as widely as possible - even if this does have a negative overall impact on innovative activity - and that deep pocket corporations have the PR savvy to understand that equating software patents and trolls/NPEs/PAEs with harm caused to SMEs and start-ups will get a much more sympathetic hearing from the media than complaints made by deep pocket corporations.

Now, I understand that I have based my argument on an estimated figure that has pretty much been plucked from the air. But the reason I have had to do that is because there is nothing else to go on – which kind of reinforces my point. It is possible that there is a serious problem with software patents and trolls/NPEs/PAEs in the US which is causing the country to miss out on the economic growth that increased innovation could provide. If there is, legislative action may well be required. But as yet, no evidence has been provided. Instead, we have numerous anecdotes, hunches and theories.

You cannot blame academics for publishing papers on what they consider to be interesting and important subjects; but you can look very closely at the agendas of entities whose interests might be served by limiting or abolishing software patent rights, as well as restricting the ability of others to sue them for patent infringement. And what is good for these entities (which by and large can quite easily live with things as they are) may well not be good for many other patent owners in the US, present and future. The fact is that before legislators start making decisions, they absolutely have to make sure that they have all the information they need to make effective ones. Right now, I would suggest, they are not even close to that point. 

Sectors

IP politics, IP litigation, Patents, IP business

Comments