Are big U.S. cities still the engines of innovation?
In a piece earlier this year on the economic problems caused by high housing costs in the city of San Francisco, The Economist’s Ryan Avent remarked, “We have machines to produce more ideas, and we call them things like ‘San Francisco’.” Avent was reflecting the large amount of economics research showing the importance of proximity and population density for generating new ideas and putting them into place. But over the past several decades, one of those new ideas, the Internet, seems to have reduce the importance of physical location. So are big cities still important idea generation machines?
A new working paper argues that while more densely populated cities will lead the way on innovation, their advantage has declined recently. The new National Bureau of Economic Research working paper tries to understand how the relationship between population density and innovation. The authors, Mikko Packalen of the University of Waterloo and Jay Bhattacharya of Stanford University, look at innovation in particular as the adoption of new ideas.
The stereotype of a great innovator is that of a loner genius who has his or her eureka moment alone in the lab. But in reality, innovation is a collaborative process that involves interplay among individuals. No innovator is an island. So in that regard, the ability to readily come into contact with other potential innovators helps boost the creation of ideas. Cities, with their high population density, are the perfect incubator for this process.
One of the ways that cities may help innovation is that they make it easy to bat around an idea or engage in debates over an idea’s merits. Gathering a group of individuals familiar with an area of research is much easier when they aren’t far apart physically. Yet the Internet’s great strength is its ability to allow people to discuss pretty much anything regardless of where they are on earth. The merits of a new idea can be debated my participants across continents. In this way, the Internet reduces the importance of population density in innovation.
To test this hypothesis, Packalen and Bhattacharya look at patents from the United States spanning from the 1870s up until 2006. Since they want to look at the adoption of new ideas, the economists focus in on the use of key idea phrases in the patents. To be more exact, they examine whether a patent is using an “idea input” that is among the top 5 percent of newest ideas. In other words, they want to see how many new inventions use new ideas.
Why should we care about the citation of new ideas? In a companion paper also released this week, Packalen and Bhattacharya find that inventions that cite new ideas are more likely to spur further innovation. New ideas, therefore, provide strong foundations for future innovation.
Their main way of looking at the effect of population density is to calculate the increase in probability of citing a new idea by going from the 50th percentile to the 95th percentile in population density. According to Packalen and Bhattacharya, the probability increase from this move between 1880 and 1910 was about 20 percentage points. By the 1970s and 1980s, the percentage point increase was only 8 points. And in the 1990s to 2000s—amid the rise of the Internet as a transformative technology—the increase was only 4 percentage points.
This new research still demonstrates that more densely populated cities foster more innovation and new ideas, just less so than in the past. Importantly, though, the role of the Internet in this shifting dynamic isn’t conclusively made in the paper. But the timing of the declines is consistent with the rise of technologies allow easier communication over long distances.
As interesting as these results are, there is one potential problem with the paper. The researchers assume that patents are a valuable metric of innovation. Given the rise of patent trolls in recent years, this assumption might not hold up for the later years under investigation.
If these results holds up under subsequent investigations, it doesn’t mean that the many economic advantages of large, dense cities are destined to fade away. Cities might not be as relatively important in the creation of new ideas, but they are still quite good at implementing those ideas. A long list of research finds that workers in more dense areas are more productive due to agglomeration effects.
Cities such as San Francisco are still idea-generating machines, but their relative importance in the generation process itself may be smaller than in the past. The world of ideas isn’t flat. But some wrinkles seem to have been ironed