Expert Commentary

Spreading the word: Geography, policy and knowledge spillovers

2012 study in the Review of Economics and Statistics looks at the diffusion of knowledge and patents across states and regions.


As U.S. states, cities and towns seek to promote economic growth by helping to foster new businesses and innovation, universities, colleges and regional knowledge centers have become a key area of focus for policy makers. Research shows that the dynamics are complex, however.

The presence of community colleges has been found to have a relationship with regional employment, but a region’s producing large numbers of graduates can have only a tenuous relationship with its human capital. Other research shows that jobs requiring analytic and social intelligence skills tend to concentrate in larger, denser metropolitan areas. This dovetails with the findings of a 2012 study that indicated that metropolitan areas with the lowest shares of creative workers tend to have higher average unemployment. Meanwhile, there are structural factors at work in the economy that appear to be limiting the number of startups and “young firms.”

A 2012 study published in The Review of Economics and Statistics, “Spreading the Word: Geography, Policy and Knowledge Spillovers,” looks at the diffusion of knowledge across states and distances. The authors, based at Duke University and London School of Economics, sought to better understand how distance and state borders affect the diffusion of knowledge from American universities. The authors refer to this dynamic as “knowledge spillovers.” The authors examine these questions using citations to university-owned patents and the extent to which corporate patents cite university scientific publications. The authors use a sample of 184 research-oriented universities in the U.S., which account for the vast majority of the nation’s research and development.

The study’s findings include:

  • Knowledge spillovers are localized, declining sharply after a distance of 150 miles. This is true for both citations to university patents and scientific publications.
  • “The probability that an inventor within 25 miles cites a university patent is 34 percentage points greater than for inventors located beyond 25 miles from the university.” After a distance of 100 miles, distance no longer has an effect on patent citation.
  • State borders constrain the diffusion of citations to university patents. Inventors in the same state as the cited university are more likely than out-of-state inventors to cite the university’s patents. The border effect is significantly stronger for citations to patents from public universities.
  • “Inventors located within the state and within 25 miles of the university are 22 percentage points more likely to cite [local patents] than inventors located at that distance but outside the state border.”
  • The impact of state borders on patent citations varies across states. The impact of borders is much larger in states that do not have non-compete labor laws. These are laws that restrict employees from moving between jobs in the same industry within the same state during a given period. The effect of borders is also larger in states with a higher percentage of inventors educated within a given state, as well as in states with a greater density and lower mobility of scientists and engineers.
  • “Moving from 0-25 miles to 25-50 miles reduces the citation [to university scientific publications] probability by 20.1 percentage points (40% of the mean citation rate), and moving out to 50-100 miles further reduces it by another 5.7 percentage points. However, citations continue to decline with distance up to about 1,000 miles.” However, the authors find no significant border effect for publications from private universities. For public universities, there is a strong state border effect for information in publications from below-average publications.

The authors conclude by suggesting further research to identify the actual methods by which distance and borders affect the diffusion of knowledge. They also emphasize the importance of understanding how this knowledge diffusion affects state economic growth.

Tags: entrepreneurship, creative class

About The Author