Category: Innovation

Scientists hold back progress by not being open enough

~ cogiddo ~ Leave a comment

Daniel Lemire, a computer science professor at the University of Quebec, has written a scathing post about scientists who care so much about their status that they do not share their data and code, even when they are obligated to, as in the case of those conducting clinical trials. Lemire’s diagnosis of the problem seems on the mark to me; his post is underlined by such fury that he has left in it some editing infelicities, doubtless created in the heat of composing it. Even researchers working in industry are better at sharing than big-shot academics, Lemire points out. Wouldn’t you be just as furious as he is on realizing this? His post is highly recommended reading.

It seems that economists need to devise a better incentive system for governing research done in universities, as well as a method for convincing the powers that be to implement these new incentives. The public shaming Lemire calls for would be a good start.

Here is a pithy quote from Lemire’s post:

The problem is that academic researchers are overly obsessed with their own personal social status. It is a stronger pull than the desire to advance science, the public good or even commercial interests.

Do patents inhibit research?

~ cogiddo ~ 1 Comment

A newly published paper by Fiona Murray, Philippe Aghion, Mathias Dewatripont, Julian Kolev, and Scott Stern asks whether patents inhibit researchers from exploring new directions for their research. The American Economic Association published a research highlight post today explaining the methods and conclusions of the paper.

Here is a quick summary: data and research tools in the sciences are increasingly covered by patents, making it costly or impossible for researchers who did not develop a set of data or a set of tools for analyzing data in their field to use these data or tools. In cancer research, many studies are conducted using genetically engineered mice. DuPont had patents on two methods for the genetic engineering mice, patents which it agreed to open up to researcher at a much lower cost than before, in two Memoranda of Understanding with the NIH in 1998 and 1999. A third method was not included in these agreements and remained as costly as it had been for researchers to access. This set up a natural experiment.

The result: mice covered by the DuPont-NIH agreements were involved in more research exploring new directions than mice that were not so covered. The authors of the paper found good ways to use citation and keyword data to capture this effect. Quoting from the research highlight post, we get to the crux of their findings:

The authors find clear evidence that the patents protecting Cre-lox and Onco limited the scope of scientific inquiry by genetics researchers, but the best policy for reducing this type of harm isn’t clear.

Acemoglu, Robinson, and Verdier ask: Can’t we all be more like Scandinavians?

~ cogiddo ~ 1 Comment

Daron Acemoglu, James Robinson, and Thierry Verdier have a new paper out that is getting a lot of attention in the economics blogosphere, rightly so. The paper can be found here. I want to make some comments of my own here, as well as drum up attention to the paper.

Acemoglu, Robinson, and Verdier present some evidence based on patents to argue that the US has been more innovative than the Scandinavian countries. They do so even when there are well-known problems with using patents to measure innovation. To their credit, after they present their theoretical model, they state that they hope it will spur more detailed empirical research. The theoretical model itself is based on a standard endogenous growth theory foundation, with some moral hazard thrown in to capture a very mainstream-economics view of innovators. Because of the moral hazard problem, innovators must be given the proper incentives or innovation will suffer.

The other crucial ingredient in their model is a spillover effect from more innovative economies to less. Their main point can be summarized as saying that economies with big social safety nets can be as innovative and wealthy these days because they enjoy innovations developed in economies with much less of a social safety net (and more income inequality) such the US’s. The development of the model is as masterful as one can expect from a group of authors that includes the author of the most advanced recent graduate-level textbook on economic growth (and Clark Medal winner in 2005, to boot).

I turn now to some of the discussion on economics blogs that I have seen in my RSS feeds, as some of the commentary is very good. First, let me start with the homo economicus assumption made in the paper (along with 99% of papers that use math-heavy economic theory, as it seems to me). It makes the protagonists, the would-be innovators, a bit too one-dimensional for believability. OK, you say, so what, if the model captures the aspect of economic arrangements it wants to capture, and does it well? But it does not, precisely because of this one-dimensionality. As a commenter (going by the name “ralmont”) to this post by Mark Thoma correctly pointed out, one of the most important innovations of the last two decades or so has been the development of Linux, on which run most of the web servers in the world, as well as the many, many phones and other devices that run Android. But Linux came out of the “cuddly” capitalism of Scandinavia (and indeed, from a then 21-year old student who opened it up to the world not in order to get rich but to learn and because he loved to tinker with operating system software). The standard one-dimensional model of economic man (the term was made up in backward times and I am keeping its gender slant on purpose, as a criticism) simply does not apply here. Also, by the nature of Linux (and all free/open source software), there is no point to going to look for a patent by which to measure the momentous impact of this development on economic growth.

In the body of the same post just discussed, Mark Thoma also says this:

An enhanced safety net — a backup if things go wrong — can give people the security they need to take a chance on pursuing an innovative idea that might die otherwise, or opening a small business. So it may be that an expanded social safety net encourages innovation.

I think I can let this stand with no further comment from me; it is so well-stated and I find it convincing.

Another blog post well worth mentioning here is one by Lane Kenworthy, a discussion of which started the Mark Thoma post I just discussed. Kenworthy makes many good points. One is that by a measure other than that of patents (the World Economic Forum’s Global Competitive Index), Scandinavian countries are almost as innovative as the US. Another is that in the 1960s and 1970s there was plenty of innovation in the US, even though that period offered less in terms of incentive to overcome the innovators’ moral hazard. A third point is that other countries with a lot of income inequality are not as innovative as as the US, so inequality does not always provide an impetus for innovation. (I can already anticipate that Acemoglu would construct a defense on this point by discussing extractive institutions, in the sense expounded in his book with Robinson Why Nations Fail, and indeed I will be looking forward to rebuttals to this criticism in the blog of the same name. I have a lot more to learn from that book and blog, which I hold in high regard—I am not writing here to bash Acemoglu and friends.) Finally, Kenworthy makes a direct comparison between the US and Sweden based on data and finds little reason to buy the story of the Acemoglu, Robinson, and Verdier paper.

Commenter “cfaman” to the Kenworthy post disputes the view of the US in the paper as not cuddly capitalism, by this:

But we are cuddly! I have a hard time thinking about innovations that have made a big difference since WWII that have not come directly our of government investment or substantially encouraged by government investment.

I have similar trouble, but then I am no economic historian.

Then there is Dylan Matthews writing that the paper’s

[…] finding relies on a model that doesn’t so much compare the American economy to social democracy as a stylized version of the American economy to actual socialism.

Well, of course, they are doing economic theory and you have to give license to theorists to simplify. But Matthews has more to say. In particular, he notes that the definition of “cuddly capitalism” in the paper

requires that there be literally no difference between successful and unsuccessful entrepreneurs’ income levels.

Here the theorist’s license to simplify may have gone too far. Scandinavian countries are not like this; in all of them successful entrepreneurs can make money, if not as much as in the US, enough to be very rich.

Matthews also offers evidence that Scandinavia is doing well compared to the US on innovation, and also points out, as the others I already mentioned did, that patent trolls and the general malaise of the US patent system makes using patent date highly suspect. Finally, he points out that even using the patent measure, the US did not become more innovative relative to Sweden countries as the US level of inequality soared in recent decades. Matthews concludes that the model of the paper does not hold because of this.

All in all, reading the debate has made me think more about innovation and for this I am thankful to Acemoglu and company. Graduate students inspired by the paper will doubtless learn a lot about how to theorize on economic growth; Acemoglu is a world-class master there. But I hope that the Acemoglu, Robinson, and Verdier plea for more empirical work on this takes precedence over theoretical refinements of the model for now. Young economists looking towards a good area for research, this is about as good as it gets. Crank up your statistical software, gather appropriate data, and go! I can imagine good ways to improve the theory in the paper, but only after it’s been bashed around a bit by the econometricians.