Economist Explains How Much Innovation Is Being Held Back: Says We Need To Fix The Patent System

from the some-metrics dept

On a recent episode of Jerry Brito's Surprisingly Free podcast, he had on economist Alex Tabarrok who recently released a Kindle short, called Launching the Innovation Renaissance: A New Path to Bring Smart Ideas to Market Fast. Part of the key argument in the book is that through bad policy, we've really held back the pace of innovation. This is something we've pointed out before, but Tabarrok has some data to back it up. One of the most frustrating things for us has been how many reports use patents as a proxy for innovation, even as many studies have shown no correlation between patents and actual innovation. In this book, Tabarrok, smartly, looks at the growth rate in Total Factor Productivity -- which is a measure of output that comes from "non-traditional" inputs -- i.e., output created directly from land or capital. This is a reasonable way of measuring the impact of technological improvement.

It's important to understand this, so let's take a step back to explain it a little more clearly. Economic growth comes from the ability to get more out of the same amount (or even better, getting more from less). Straight output increases can come just from increasing capital or increasing labor, but you're not expanding the overall pie. Actually creating economic growth -- providing more than was possible before -- comes from knowledge and ideas being applied to various problems. Knowledge and ideas can effectively create something from nothing, or create something more from something less. Greater and greater efficiency can be driven by smart ideas over time (mass production, automation, computerization, etc.), allowing for economic growth and the ability to create some amazing new things and expand the economy.

Total Factor Productivity is effectively measuring this point: how much are we getting from the factors of production. Over time, Total Factor Productivity should continue to rise, as we improve efficiency and create new things and generally grow the overall pie. But it's the rate of growth that is most important in determining if we're growing at an optimal pace. If the rate of growth of TFP is slowing, then that suggests something is getting in the way: and that's exactly what Tabarrok is talking about. Over the last few decades, rather than expanding, the growth rate for TFP has been declining. Innovation is still happening, and we're still growing the economy, but not at the same pace as in the past, and that's a problem, since normally we should see the rate of growth increase as well.

In short, as Tabarrok points out, we're seeing important growth and innovation happening later than it likely would have otherwise, if current policies weren't holding back economic advancement. He notes that if the pace of growth in TFP had continued at the rate it had in the two decades after World War II, we'd be living in a world of innovations from 2076, rather than what we have today. In other words, the pace of innovation has slowed down so much in the last 40 years, that it will take another 65 years or so just to get to where we should be today. That's terrible.

Among the things to blame? Obviously: the patent system is at the top of the list. We've certainly been down this road before, but Tabarrok hits on many of the points we have concerning how patents hinder the pace of innovation. As you may recall, a few months ago he actually put together a nice video on the topic as well. The key point: patents serve to make follow-on innovation much more expensive, and follow-on innovation is often key to continued productivity gains.

Brito asks him about one of the key retorts we often hear, noting that we do see plenty of innovation, and even the areas where there are the biggest patent fights, such as the smartphone patent wars, involve a bunch of companies -- mainly (though not entirely) in the US, leading the innovation. Tabarrok points out that, again, the real issue here is the rate of growth, not the fact that there is innovation happening:

You have to think about where is innovation most likely to be slowed by the patents. And it's actually in a field in which the natural rate is quite high. Suppose we have a field where the natural rate of progress is low and something is patented. Well, since the rate of progress is low, you're not raising the cost to future innovators very much, because they were going to wait anyway, because there wasn't much to build upon. So industries where you're not building on top of other innovations, where you don't have cumulative innovation, then a strong patent is not going to do very much. It's not going to slow things down.

It's precisely the areas in which there's lots of innovation, where you're building on previous innovations and things are moving rapidly, that the patent does slow things down. So, yes, it's true that the smartphone industry is innovating, rapidly. But it's my belief that they're innovating less rapidly than they otherwise would. And it's precisely in fields that are innovating rapidly that you expect patents to have the worst effects.

This a key point that is all too frequently misunderstood by patent system supporters. Even in the US government, where we've pointed out that they use incredibly simple claims of correlation, like Steve Jobs getting patents as proof of a patent system working. But that has nothing to do with the pace of innovation. It's that second-order number that is what is actually important in determining if we have the most effective policy. It's honestly tragic how few people understand the importance in understanding the rate of change in something, rather than its absolute rate in looking at the impact, and it's that failure to understand this (don't we teach calculus any more?) that has resulted in some really bad policy decisions.

Tabarrok discusses the patent nuclear war in the smartphone space and points out, as we have in the past, that it has nothing to do with innovation at all, but is just moving money around, mostly to lawyers. Even worse, he notes that the truly disruptive and breakthrough innovations tend to come from outside the big mainstream players, but rather from small upstarts coming out of left field -- but in a heavily patented area, those players don't stand a chance, and those breakthrough, epoch-defining innovations that move the entire state of the art vastly forward, die an untimely death, often before ever seeing the light of day. As he notes, those innovations get killed in a court of law, rather in the marketplace, which is terrible for innovation.

In terms of solutions to this issue, Tabarrok suggests getting rid of software patents first of all. Also, he suggests different patent lengths for different types of inventions, with shorter inventions for those which cost less to develop. I've seen this suggestion made before, and I'm not against it, though I do wonder about the unintended consequences there. If you're providing patent lengths based solely on "cost of development," you've now created perverse incentives for inventors to spend more on invention than might otherwise be necessary, as they attempt to get the longest possible patent. We should want the invention process to be efficient as well, including seeking to minimize the cost of development where possible. Developing things cheaper is, in fact, one key area of innovation. So while I appreciate the concept behind the idea, I'd be worried about such a system having significant consequences that might, in fact, slow down innovation even further.

Tabarrok does, however, suggest an alternative way of determining the length of a patent that might avoid some of the unintended consequences above: letting companies self-select, based on certain trade-offs. For example, he suggests changing the level of scrutiny and the time to get a patent based on what length patent you're seeking. So, if you want a 20 year patent, then it's going to get tremendous scrutiny, a careful look at the prior art and a detailed look into whether the invention passes the non-obviousness test. However, if you're willing to just take a three year patent, then the standard of review is much lower, and the patent is granted much more quickly. For those innovating in rapidly changing fields, you could see them choosing this option, knowing that they'll get the patent faster, and that by the time the patent expires, the world will likely have moved on already.

Brito also makes what I believe is a mistake in suggesting (as Tabarrok implies) that where the cost of investment is higher, you necessarily need greater inducement to do the investment. Again, I appreciate where both are coming from, but I believe this is based on a false premise, that in order to innovate, innovators need a special extra inducement to invest. Studies have shown that this is rarely the case. Inventors invent most frequently not because of the inducement of a patent system, but (a) because they need the invention themselves and (b) because they see a market need, and recognize they can recoup the investment and more in the marketplace through selling a good product. Starting from the place of believing that innovators need "inducement" when the market and basic needs may already provide such incentives seems to fall into the trap of believing what patent system supporters always claim: that the patent system is necessary for such inventions and innovations.

This post is long enough, but the podcast (and the book) goes on to talk about some other policy issues that are important for encouraging greater innovation: fixing our incredibly broken education system and also fixing our broken immigration policy towards high-skilled labor. These are both issues that we've discussed around here as well, and Tabarroks' views are quite interesting, and worth checking out.

Filed Under: alex tabarrok, innovation, pace of innovation, patents, total factor productivity