The Benefits of City Locations

Process improvements don’t grow on trees, say three business school professors, but they might be native to cities.

Take productivity-boosting intranets, for example, the private Internet capabilities that enable companies to improve information sharing and collaboration within a company. Developing such projects often requires a substantial investment of resources, largely devoted to programming talent. One source of programmers, of course, is a company’s information technology staff, whether from the particular office or transferred from other field offices.

When talent isn’t available inside the company, however, cities may act as a substitute, the three professors found. “The primary reason is the thick labor markets inside major cities,” explains Shane Greenstein, Elinor and Wendell Hobbs Professor of Management and Strategy at Northwestern University’s Kellogg School of Management.

Greenstein, along with Chris Forman, assistant professor of IT management at Georgia Tech’s College of Management, and Avi Goldfarb, assistant professor of marketing at the University of Toronto’s Joseph L. Rotman School of Management, examined company- and office-level survey data from 1998 through 2000, from Harte-Hanks Market Intelligence’s Computer Intelligence Technology database. In all, they covered 86,879 offices of 100 or more employees, belonging to 47,966 different companies.

The 2007 working paper Understanding the Inputs Into Innovation: Do Cities Substitute for Internal Firm Resources? — in press at the Journal of Economics and Management Strategy — explains how the authors looked at rates of investment in process innovation (or at least in intranets), based on company and office programming staff and office location. Sure enough, offices in urban areas — in metropolitan statistical areas with populations over 500,000 — were 3.26 percentage points more likely to make such investments, suggesting the city’s process innovation benefits.

Further reasoning that large companies — those with multiple offices — had more resources to implement cutting-edge technologies, the authors expected to find higher rates of investment for field offices of larger companies than for single-office companies. “We were sure that the big firms move programmers around,” explains Greenstein. “They have pilot projects in one location and then re-implement them elsewhere.” The data confirmed that to be true — but only just. “It looks like single establishments in urban areas are not at much of a disadvantage. That was an interesting thing we found,” says Greenstein. “[Single-office companies] seem to be able to go to their labor markets and achieve an outcome that looks similar to the multi-establishment firms.” When programmers weren’t available inside the company, the company simply outsourced. It’s the process analog to the classic make-or-buy decision in manufacturing.

But rural companies may not have the option to buy. “Single establishments in rural areas are at a disadvantage, particularly for this technology,” says Greenstein.

As expected, the rural penalty was less noticeable for rural offices belonging to large companies. “Clearly, being part of a large organization helps and it partially overcomes isolation,” Greenstein says.

What this tells companies is that cities and their creative classes are attractive locations because of their ability to feed innovation — higher cost of living notwithstanding. The larger populations of urban areas make it easier to invest in discrete process improvements. In fact, the authors found that cities with a greater population of IT workers were associated with an increased likelihood of intranet investment, cementing the relationship between urban work forces and innovation. Although larger companies may care less, because they have the resources to initiate process improvements on their own, smaller enterprises might consider the innovative potential of the neighborhoods in which they open offices.

But urban benefits extend only “to the extent that [innovation] involves thick labor markets that sit proportionately in urban areas,” Greenstein says. It might not be relevant for, say, cutting-edge tractor technology or numerically controlled machine tools, he says, because labor markets for skilled technicians are spotty and not necessarily diffuse throughout different populations, unlike programmers.

In addition, product innovation may require internal resources and local talent. In research-and-development-intensive industries, like pharmaceuticals, for example, Greenstein explains that “it really does help to be sitting next to another pharmaceutical company. You end up sharing a labor market and sharing technical talent. What that effectively means is that for R&D-intensive output, your internal resources are a complement, not a substitute, to your location. You might think there’s some substitution, but it would be less binding than with process [innovation].”

For more information, download the paper or contact Chris Forman at chris.forman@mgt.gatech.edu, Avi Goldfarb at agoldfarb@rotman.utoronto.ca or Shane Greenstein at greenstein@kellogg.northwestern.edu.