Silicon Valley. Boston. New York. In the U.S., innovation is becoming increasingly concentrated. Between 2015 and 2019, eight mostly coastal “superstar” metro areas accounted for nearly half of the nation’s technology sector job creation, according to a report from Brookings Metro. Scientists, engineers, and entrepreneurs flock to their peers — to the places where invention is in the air.
But there is a downside to this density of talent. The most technologically productive places in the country also have some of the highest labor and real estate costs. Startups deciding where to locate as well as established companies opening new offices must actively weigh the benefits of productivity in a given location against the costs of doing business there.
In a new working paper that examines the balance between local productivity and local costs in research and development, MIT professorsand Jonathan Gruber and University of California, Berkeley’s Enrico Moretti find that the productivity gains from a density of scientific talent generally outweigh the additional costs.
At the uppermost margin, however, which is to say in cities with the greatest concentration of talent, productivity gains are essentially washed away by dramatically increased labor and real estate costs, with real estate costs accounting for most of this result.
“This is good news for, say, Kendall Square, in Cambridge, Massachusetts, which has lots of scientists,” Johnson said. “But not if real estate prices rise to San Francisco levels. At that point, the cost increases outweigh productivity gains, at least for standard research and development.”
The researchers’ work builds on a 2021 paper by Moretti in which he quantified the advantage of clustering talent in one place. He found that scientists who live in areas with 10% more colleagues in their given field produce 0.5% – 0.9% more patents per year.
Pricey, but worth it
Using those findings as a springboard, the researchers collected average wage and real estate costs associated with research and development in 133 U.S. cities between 2010 and 2015. The researchers used Glassdoor to capture wages, and CoStar and the American Community Survey to assess land values and rental expenses.
Comparing the productivity gains in locations densely populated by scientists against the wage premiums and higher real estate costs, the researchers find that productivity gains tend to outweigh additional costs by a factor of six. Specifically, a 10% increase in the scientific population is, on average, associated with a 0.11% increase in total costs and a 0.67% increase in patent creation.
In short, firms moving from cities with a small population of scientists to cities with a large population of scientists might expect to experience productivity gains up to six times larger than the increase in production costs.
Importantly, though, the researchers find diminishing returns in areas with the highest density of scientists. In these places — the Bay Area, Boston, and Honolulu, for instance — the increase in research and development costs counterbalances productivity gains.
A role for the federal government
These findings hold particular relevance as the federal government redefines its role as an investor in innovation. Johnson cited three recently proposed pieces of legislation:
- The bipartisan Endless Frontiers Act, which would commit $10 billion over the next five years for grants to create 10 – 12 new technology hubs.
- The Innovation Centers Acceleration Act, which would provide $80 billion over ten years for a competition for cities to become technology centers.
- The Federal Institute of Technology Act, which would invest nearly $1 trillion in public R&D over 10 years, targeting technology centers with a significant share of the funds.
These investments have the potential to dramatically accelerate research and development around the country, Johnson said. But properly evaluating the effectiveness of such place-based polices requires understanding and carefully measuring both costs and benefits.
“Government money and private investments have been pulled towards the higher-density places on the East and West coasts for several decades,” Johnson said. Medium-density places, like Rochester, New York or Pittsburgh, Pennsylvania, deserve more attention, he said.
What’s more, building lots of mid-sized hubs for innovation would not only be good economics — there are lots of positive effects and social gains that flow from knowledge creation — but also good politics.
“Innovation is a central part of our modern economy; we need that innovation to take place in all parts of the country, not just in a few places,” Johnson said. “It’s fine to have talent magnets in specific places, but if this is all you have, then everyone else becomes frustrated.”