by Chris Forman, Avi Goldfarb and Shane Greenstein, Voxeu.org
The diffusion of the internet has had varying effects on the location of economic activity, leading to both increases and decreases in geographic concentration. This column presents evidence that the internet worked against increasing concentration in invention. This relationship is particularly strong for inventions with more than one inventor, and when inventors live in different cities.
Reading the technology press, it often seems as if the media think all high-tech invention happens in Silicon Valley. This parochial viewpoint highlights the ‘agglomeration’ advantages that the Valley provides to inventors because so many technology firms are located in the same place. These advantages include easier access to funding from local venture capitalists, sharing of fixed costs such as specialised patent lawyers, and easier exchange of ideas between researchers.
Placing such emphasis on the power of agglomeration runs counter to the ‘flat world’ and ‘death of distance’ predictions of Thomas Friedman (2005) and Frances Cairncross (1997). These writers argued that the internet would eliminate geographic barriers to economic (and political) activity.
Cairncross’s and Friedman’s predictions were informed by clear economic trends. Declines in the costs of coordinating across long distances enabled the disintegration of supply chains on a global scale. Declines in communication costs also made it easier to collaborate over long distances, leading to changes in the geographic distribution of innovation and invention (Agrawal and Goldfarb 2008, Forman and van Zeebroeck 2012).
This difference between the reporting in the technology press and the ideas of Friedman and Cairncross suggests a more general question: Has the geographic concentration of inventive activity become more or less geographically concentrated? Or, simply, which forces matter most?
The internet’s effect on the geographic concentration of invention
Enough time has passed to move beyond observation and anecdote. It is possible to collect statistical evidence about the geographic spread of invention, and analyse its change after the Internet deployed.
Our research looked specifically at whether or not the deployment and adoption of the internet increased or decreased the concentration of inventive activity in the US between the early 1990s and the early 2000s. This timeframe is significant because the commercial internet began its explosive growth in the US in approximately 1995. Because the internet did not become intensively used in every business and in every location at the same time, this time period also permits us to examine whether the internet’s diffusion played a role.
To be precise, by increase in concentration, we mean that the places that served as the location for the majority of inventions in the past served as the source for an even greater share in the future. By decrease, we mean the opposite – that the places that are not rich with invention become richer. We measure invention using data on patents.
Here is the bottom line:
- The diffusion of the internet worked against the trend toward increasing geographic concentration of inventive activity.
Upon a first glance at the data, this conclusion is not obvious. Our evidence consists of the growth rate of patenting across counties. Patenting grew by 27% during this period. For the top quartile of patenting counties from 1990–95, patenting grew by 50%. For those below the median, patenting did not grow at all. So, overall, location became more important over time.
This overall trend is driven by substantial increases in the concentration of patenting at the very top of the distribution, such as in Silicon Valley (e.g. Santa Clara County, San Mateo County, and San Francisco County). In other words, the trend at the very top is driven by a small number of places.
Looking at the broad trend in the rest of the US, however, we showed that many places resisted the overall concentration of invention. To understand this, it is important to separate the adoption of the internet (and associated advanced technologies for digital communication) by businesses from the invention of new technologies.
The internet reduced the importance of Silicon Valley for invention
We show that heavy use of the internet mitigated the trend toward increasing concentration of invention. This relationship is stronger for inventions with more than one inventor, and especially strong for inventions where the inventors lived in different cities. In other words, by enabling firms outside of Silicon Valley to communicate with researchers elsewhere, the internet reduced the importance of the Valley (and other tech innovation centres such as Austin, Texas).
Thus, Cairncross and Friedman were correct that the internet reduced the importance of location for invention; at the same time, however, other forces were increasing the importance of Silicon Valley, Austin, and a handful of other places.
Why does this matter? Most prior work examining economic activity other than invention finds that the forces pushing toward more concentration tended to win after the commercial Internet diffused. The effective use of advanced internet technology draws on frontier IT skills that are found disproportionately in urban areas. Sophisticated internet business processes also build on existing links between business use of IT, support services, and specialised labour markets in urban areas. Furthermore, while the internet reduces communication costs, most communication and most social contacts are local (e.g. Wellman 2001). Therefore the impact of the internet on wages, productivity, and other measures of economic activity seems highest in leading urban areas (e.g. Forman, Goldfarb, and Greenstein 2012).
However, invention shows a different pattern. The internet appears to be weakening the links between the geography of inventive activity and the geography of other economic activity. Such a broad trend, if sustained for a long time period, could result in numerous changes to the geographic concentration of different parts of the value chain. So while the technology press may be right that Silicon Valley accounts for a disproportionate share of invention, the information and communication technologies coming out of the Valley have reduced the importance of technology clusters like Silicon Valley to the overall rate of US invention.
In closing, we must add a word of caution, and this is directed at the applied econometricians among our readers. (You know who you are!) We cannot rule out the possibility that an omitted factor caused both business use of the internet and growth in patenting in the set of internet-adopting counties in the 1990s. And to you we say: if you care that much about the causality, then don’t stop with a column. Go and read our paper (Forman et al. 2014). More to the point, there is a lot to be done on this topic, and one paper can only be one step towards a definitive answer. We welcome additional insight, and encourage additional research into the issues.
- Agrawal, Ajay, and Avi Goldfarb (2008),”Restructuring Research: Communication Costs and the Democratization of University Innovation”, American Economic Review, 98(4): 1578–1590.
- Cairncross, Frances (1997), The Death of Distance, Cambridge, MA: Harvard University Press.
- Forman, Chris, Avi Goldfarb, and Shane Greenstein (2012), “The Internet and Local Wages: A Puzzle”, American Economic Review, 102(1): 556-575.
- Forman, Chris, Avi Goldfarb, and Shane Greenstein (2014), “Information Technology and the Distribution of Inventive Activity”, NBER Working Paper 20036, forthcoming in Adam Jaffe and Benjamin Jones (eds.), The Changing Frontier: Rethinking Science and Innovation Policy, University of Chicago Press.
- Forman, Chris and Nicholas van Zeebroeck (2012), “From wires to partners: How the Internet has fostered R&D collaborations within firms”, Management Science, 58(8): 1549–1568.
- Friedman, Thomas L (2005), The World is Flat: A Brief History of the Twenty-First Century, New York, NY: Farrar, Straus, and Giroux.
- Wellman, Barry (2001), “Computer Networks As Social Networks”, Science, 29: 2031–2034.