Other people's business

RSS

Should the White House crack down on patent trolls, or on patents?

That patents boost innovation is received wisdom, but it might not be true.

A young Albert Einstein. The physicist was a patent clerk as a young man. Photograph: Getty Images

The Obama administration has sung the tune that Silicon Valley wants to hear, and will be cracking down on patent trolling – even announcing the policy in the web-friendly format of animated gif:

Patent trolling is the colloquial name for a bundle of strategies used by companies technically known as "non-practicing entities", which own patents on technologies which they have not brought to market. The patents used are frequently broad claims, granted by patent clerks with a lack of knowledge of the area. Often they stem from the early days of the internet, and cover ideas which can be summed up as "use the internet to do something".

Once the NPE has got hold of the patent, usually through buying up the portfolio of an older company, they then start threatening smaller companies with lawsuits. Key to the plan is that few of these lawsuits actually come to pass – instead, the companies are bullied into paying "license fees" to avoid the suit. That's because the trolls could ill-afford to have too many cases going on at once, but also because once they head to court, they run the risk that their patent will be invalidated (or at least that it will be ruled to be much narrower than they've construed it).

The "non-practicing entity" part of the equation is most important because many industries exist in a tangled web of patent infringement and licensing. Apple probably infringes a bunch of Google's patents on something – just because Google holds so many patents – but the converse is also true. The two companies end up existing in a state of mutually assured destruction, which NPEs can bypass.

But the other reason why NPEs are the focus of the changes is that they're easy to vilify. They don't make useful products, they don't have public profiles, and many of them don't even have any real link to the person who filed the patent in the first. But they are merely a symptom, not a cause, of the problems of the patent system.

Overly-broad patents have been granted to practicing entities as well. Amazon managed to patent the idea of ordering an item with one click, and it took eight years and a whole lot of publicity for the US patent office to force them to amend it. And they aren't just granted in technology. Here's a patent for "the process and apparatus for refreshing bread products, particularly open face items such as sliced rolls, buns, muffins, and the like" by heating them. You may know it as a "toaster". The patent was granted in 1999 and only lapsed when the inventor stopped paying the renewal fees.

And while it's easy to vilify NPEs, they do have a reason for existing. Theoretically, patents encourage two things: invention, and disclosure. Because you can't patent something without fully describing what it is and how you do it, it ensures that creators reveal their inventions to the world, so that when the patent lapses, everyone else can copy it. In theory, NPEs encourage inventors who might not have the wherewithal to bring something to market to still disclose it; and then, with the simple payment of a license fee, others can build on it.

In practice, of course, this doesn't happen. The industries where patent trolling is most common have little need for this sort of disclosure, since they're ones where it's encouraged in other ways; biotech leans heavily on peer-review, and coding on open-source. And the sort of actually-insightful inventions which we want to be disclosed are useless for patent trolling, since by definition they won't be independently invented. Instead, patent trolls need patents which as many people will infringe as possible.

But if we're comparing theory to practice, we need to go further than just asking whether patent trolls are broken in practice. What about the entire system?

Patently problematic

Patents have been around for so long that it's often forgotten that, prima facie, they're a bad thing. A patent is a government-granted monopoly on the exploitation of an invention, limited for a short time period. The immediate effect of that is the same as with all monopolies: the monopolist gets the chance to limit the supply of the product in an effort to increase their profit margins.

So patents start off on the back foot, and must justify their existence somehow. That justification comes back to the theoretical advantages of patents encourage: invention, and disclosure. The idea is that the monopoly, and the increased money which the inventor accrues from it, encourages innovation. Without patents, it is claimed, inventors would not be able to make any money from their creations, and may just not bother.

Even that advantage is actually a post-hoc justification, though. In the early days of patents, before their commission was codified and organised, they were granted as a sort of reward to inventors. The economic concepts of incentives had not yet come about, and the best way to get a patent was to have a mate in parliament who could make the case that a state-granted monopoly was your just desert.

That system was the worst of both worlds. Since a patent wasn't guaranteed to an inventor, it could do little to encourage invention; but if it was granted after the fact, it would do a great deal to stymie growth in that sector in the future. Economists Michele Boldrin and David Levine, in their book Against Intellectual Monopoly (available online in its entirety, naturally), detail what happened to the steam industry in the years during and after James Watt patented his engine:

During the period of Watt’s patents the U.K. added about 750 horsepower of steam engines per year. In the thirty years following Watt’s patents, additional horsepower was added at a rate of more than 4,000 per year. Moreover, the fuel efficiency of steam engines changed little during the period of Watt’s patent; while between 1810 and 1835 it is estimated to have increased by a factor of five.

Patents do harm. The question is whether they do good as well, and enough good to outweigh the harm

As the system has progressed since then, the scope of what is patentable has increased dramatically. Whereas the patent system initially covered physical inventions only, it soon grew to encompass things like business methods, software innovations, and genetics. If patents encourage innovation, then we would expect a burst of invention shortly after they're expanded to a new area. That is not quite what Boldrin and Levine find.

They look at a number of examples, from US agriculture – which grew from a system where only "some mechanical and chemical inventions" could be patented to one where plants and biotech inventions could be – to the software industry. In every case, they find that the extension of patents to the industry had little effect, or even a negative effect, on productivity growth.

It's hard to measure the actual effect of patents. Partially that's because the typical measure for the level of innovation in society is the number of patents granted – which is obviously not suitable for an inward-looking examination of the system. But it's also because their existence is such received wisdom that we have to hunt hard to find examples where they aren't in effect. And many of those examples are new fields, where we'd expect a burst of innovation anyway.

Nonetheless, the balance of evidence isn't as clear cut as popular perception would have it. If the Obama White House really wants to help America's innovators, and free them from the tyranny of being sued for invention, it could consider ending patents entirely.