United Nations University-Maastricht, Université Louis Pasteur, French Ministry of Research and Education
NAME: Robin Cowan
TITLE: Professorial fellow, United Nations University-Maastricht Economic and Social Research and Training Center on Innovation and Technology; professor of economics, Bureau d'Economie Theorique et Appliquée at the Université Louis Pasteur; and professor of economics of technical change, University of Maastricht; Chaires d'Excellence, Ministry of Research and Education (France)
BACKGROUND: Assistant professor, economics, University of Western Ontario; PhD, economics, Stanford University
It has long been hypothesized that the presence of one or more research universities in a given region is associated with increased industrial innovation and economic development in that same region. However, the underlying reasons for this “university effect” have yet to be pinpointed.
Since the inception of the Bayh-Dole Act in the US in the early 1980s, university technology transfer has been thought to be the main driver of the university effect through patenting and licensing, the publication of research papers, and the establishment of university spinout companies.
However, according to a recent working paper published by researchers affiliated with the United Nations University-Maastricht Economic and Social Research and Training Center on Innovation and Technology (UNU-MERIT), technology commercialization activities may not have as significant an impact on the university effect as previously thought.
According to the study, which examined how new science, medicine, and engineering departments or universities in Italy affected regional innovation from 1985 to 2000, the creation of a new department or university on average increased regional innovation activity within three to four years, and led to a 10-percent increase in the number of patents filed by regional firms.
But the researchers also found that traditional measures of academic research and commercialization activity could only account for about 30 percent of this effect.
A full version of the study can be seen here. This week, lead author Robin Cowan took a few moments to discuss with BTW the results of the study and its implications for the field of technology transfer.
What was your motivation for this study?
There is a lot of interest now in the role of universities in the larger innovation system of an economy. Over the last couple of decades, we have seen the universities being pushed towards the market, and being expected to provide close to marketable innovations for industry to pick up, take to market, and make profits on.
Do you think this increased interest is worldwide? It seems like there has been an increased interest in this in the US for a decade or more.
It’s probably closer to two or three decades. It’s spreading around the world. It started in the English-language world, and it’s spreading.
Who paid for the study?
The funding for my postdoctoral fellow, Natalia Zinovyeva, came from Chair d’ Excellence, which I hold in France. Most of the funding comes from our salaries, because most of the input was our time.
Why did you choose to study the Italian university system?
Basically for statistical reasons. When you’re doing a study like this, there is a big endogeneity problem. Universities open, and then have an affect on what happens in industry. That’s the general idea that we’re interested in. But why do universities open? Maybe they open because industry is vibrant in an area. So when you think you’re seeing universities affecting industry, what you might actually be seeing is vibrant industry attracting universities. You also get the causation backwards. It’s very difficult to sort this out statistically.
The reason we looked at the Italian case is that it seemed, over the couple of decades that we looked at this, that this wasn’t the case. Analyses of these new universities suggested that they were opened more or less at random.
This is what you referred to in the paper as a ‘raindrop pattern?’
Did you primarily consider the openings of new universities, new departments in universities, or both?
‘Faculties’ is the way they are described – new faculties in a region. Sometimes it was whole new universities, although not very often. But often it was a new faculty at an existing university, or some kind of branch operation of a university, where it opened a second campus 30 miles away. But the interest for us was the actual faculties, because very often these were the case of a university opening a new faculty, such as chemistry.
A large portion of the new faculties that you studied were in the science, engineering, and medical fields. Did you get a sense that one of these areas exerted more influence than others on regional industrial innovation?
We didn’t. In a study like this, there are many things that affect the dependent variable. We picked out a few. From our point of view, things tend to be hard to identify statistically. I guess things like medicine and chemistry probably have a bigger effect than other industries, but we can’t identify that statistically.
Your major finding was that about 30 percent of the ‘university effect’ on regional industrial innovation can be attributed solely to new applied and basic research, patenting, and resulting publications.
Yes. Here’s the sort of stripped-down version. You open a university or a faculty, and a few years later, you see more [industrial] patenting in the region. You could say that maybe it was because there are more university graduates. But we shortened the time frame to eliminate that effect. Then you might say that it’s because university scientists are taking out a lot of patents, which then leads to industrial patents; in other words, a university scientist will patent something, and industry will pick it up and create the extra patents it needs to turn into a product with surrounding pieces of technology. Or maybe it comes from the fact that because there is a university there, industry is more aware of what’s going on in the basic sciences.
We saw three channels that we could identify, and these are the channels that people always talk about, such as university patenting, spinoffs, and those sorts of things. And we asked, ‘To what extent can the total effect of a new faculty on industrial activity be ascribed to those three channels everyone talks about?’ And what we found was that of the effect of opening a faculty in a region, only 30 percent can be ascribed to those channels. So 70 percent of the effect of a university in a region is not ascribable to those channels. It has to be something else.
So it is well-documented that regional industrial activity spikes in response to new faculties, but it is unclear what the contributing factors are?
That’s right. We’re not saying that universities have no effect on industrial activity – in fact, we’re saying that they do. The idea is that a large chunk of the effect is not due to what people are talking about.
How do your findings fit with the general school of thought that technology development and tech transfer greatly affect regional industrial activity?
Our findings are sort of swimming against the tide, which is part of the reason we did the paper. We’ve had a couple of decades of thinking that universities are full of patentable things, that we should get university professors to patent all kinds of stuff, and then industry will really take off – because the way universities serve the innovation system is by producing a whole bunch of almost finished products or processes.
One of the reasons people started thinking this was because of biotech, because it is kind of true in biotech. But the argument I’ve been making for a long time is that even if it’s true in biotech, you can’t generalize outside of biotech. Biotech is weird in a lot of ways. The thing that has been happening recently and the thing that motivated me to do this paper is that universities have been remodeling themselves on the biotech model, and treating all disciplines and sectors as if they could be or should be like biotech. Obviously this is not the case. Tech-transfer people might not like the findings of this paper.
Tech-transfer people are doing their job and trying to sell the technology that is developed at the university. In a way, you can’t fault that provided it doesn’t distort the activity of the university. But we see evidence in some places that it does. For example, in Denmark, I have been recently told that the amount of funding a university receives is based on many things, but one of them is how many patents it owns. That is a very big change in what universities think they should be doing. Then the government is explicitly saying that you should be patenting things, and we will reward you for patenting these things. A lot of people I talk to are sort of horrified by this notion.
Do you think these findings would be replicated in the US?
Yes, I do.
Are there any geographical constraints to consider? For instance, maybe a large portion of Italian academic patents get translated into industrial development and innovation in Italy only, as opposed to patents coming out of, say, academic institutions in California, which might spur industrial innovation nationwide or internationally?
You’re right when you say that different innovation systems work in different ways, in the way you described. I haven’t given a lot of thought to that, but my guess is that it wouldn’t affect the main message. We’re not seeing this because all industrial innovation is taking place in Rome. I don’t think that is the explanation. If you look at a particular region, we see a university appear, and a couple of years later, we see more industrial activity. So that kind of mismatch between where the university is and where the resultant innovation is has been taken care of statistically, in this case. We see a university open in Umbria, and three years later we see a 10 percent increase in patenting in Umbria. How much of that 10 percent can we ascribe to these three channels discussed in the paper?
Do you plan to study this elsewhere?
We would love to. But as I said, there is a really big endogeneity problem, and the reason we did this is because of the unique situation in Italy. If we had a way of solving this endogeneity problem somewhere else, then for sure we would do it, because as you’re hinting, it would be nice if this effect were obviously more general than just a gut feeling that it was more general.
You leave open what might be contributing to the other 70 percent of the university effect on industrial innovation. Have you thought about what else might be contributing to the effect?
When anybody answers that question, it sounds like a ‘castles in the air’ kind of answer. But I think that one of the things a university does is provide outlets for people with certain kinds of creativity, and an environment in which those people are happy – and I don’t mean just inside the university buildings, but within the general surroundings of the university. I run the risk here of sounding like an elitist. But I think that having a university around provides a group of people who have certain ideas about innovation, and thinking new thoughts, et cetera, and this can only be conducive to more of the same.
And this goes beyond the effect that recent university graduates might have because you saw these industrial innovation spikes within a few years of a new faculty being opened, correct?
Yes. One of things that I’ve been told — and this is purely anecdotal — is that big firms find that within their companies there is a fairly small proportion of very creative people who do a lot for the firm’s innovation. And all firms, but especially big firms, try to find out what those people are like and how to make their environment attractive.
One of the things those people like to do is university-type activities – I don’t mean taking classes, necessarily, but going to seminars and talking to graduate students, et cetera, which is much easier when there are universities around. This may make me sound like what all universities do is sweetness and light, which I don’t think. But I think there is a big component of this kind of thing which is extremely difficult to pin down. And I think that part of the reason we see a big push on universities to move toward the marketplace is that you can put numbers on things like how many patents you own, and you can’t put numbers on these other things.