The New Geography of Innovation

Jobs don't follow talent. Jobs follow knowledge production.
Publish date:
Social count:
Jobs don't follow talent. Jobs follow knowledge production.
Margaret Morrison Carnegie Hall at Carnegie Mellon University. (Photo: Tianming Chen/Flickr)

Margaret Morrison Carnegie Hall at Carnegie Mellon University. (Photo: Tianming Chen/Flickr)

Recent news of Uber locating its self-driving car research efforts in Pittsburgh sums up the economic geography as talent "poaching." If that is indeed the case, then why didn't Uber entice the brains to San Francisco? Surely the Bay Area has more allure than Shittsburgh. Indeed, Uber raided Carnegie Mellon University's National Robotics Engineering Center:

One likely driving factor in NREC departures is money. Uber is flush with it, with recent reports suggesting the company is raising another $1.5 billion in funding, valuing the company at $50 billion. As part of this deal, Uber paid CMU an undisclosed sum to fund faculty chairs and graduate fellowships. CMU declined to comment on its partnership with Uber, including questions about how much money changed hands and who owned the resulting technologies.

Yes, much talent will be leaving the NREC. But instead of moving across the country or just out of state (something Pittsburgh is used to enduring), Uber will set up shop just a mile down the Allegheny River. Why? People migrate. Anchor institutions such as CMU do not. "Flush" with cash, Uber would have no problem pulling prospective employees to wherever the company saw fit to house them. The site location decision is curious.

More curious is the "undisclosed sum" paid to CMU. Apparently, the university has some leverage in this transaction. Uber can't take the knowledge and run. But venture capital usually dictates where companies must go. Something has changed. The economic geography has shifted. Uber must come to CMU, not stay put in San Francisco. Missing from the picture:

While the prior research has shown the importance of academic research to the development of specific local industries, such as pharmaceuticals or electrical and electronic equipment, and that the productivity gains from academic research tend to be highly localized,we still have little understanding of the causal role that research university activities play in contributing to broad-based regional economic development or the extent to which they facilitate knowledge-based agglomeration.

Academic research could inform where industry locates. Conservatively, "empirical results indicate a causal link between university research activities and productivity gains in neighboring firms." The authors call for more inquiry. Richey Piiparinen and I have responded.

In "From Metal to Minds: Economic Restructuring in the Rust Belt," we detail the transformation of Pittsburgh's regional economy from manufacturing to knowledge production (thus explaining why Uber would situate itself in close proximity to Carnegie Mellon University). We were surprised to discover, however, that Southwestern Pennsylvania failed to promote the tech industry. The findings are ironic in light of strongly positive indicators. If Pittsburgh didn't succeed in economic restructuring, then why did metrics such as per capita income and educational attainment pop so impressively?

We noted a similar trend in Boston, Waterloo (Ontario, Canada), and Cleveland. Wherever world-class research was done at a university, private firms (e.g. Google) would cluster like moths to a flame. The quality of knowledge production acts as a rationale for economic geography.

Jim Russell, a geographer studying the relationship between migration and economic development, writes regularly for Pacific Standard.