(Bloomberg Opinion) -- As a few technology hubs and big cities leave much of the country in the dust, towns and urban areas across the U.S. are emptying out and falling into decay. Simply helping people in the left-behind regions move isn't a satisfactory solution; flourishing cities are often unwilling or unable to accommodate large influxes of new residents, and half-empty towns can’t sustain themselves with a shrunken tax base. Instead, the government needs to do something to revive struggling regions.
In recent years, some thinkers have pushed the idea of using research and development spending to encourage industrial clustering. Evidence strongly shows that government-backed R&D is good for businesses. So why not just direct that spending toward areas that need new business the most? That was the idea put forward by Massachusetts Institute of Technology economists Jonathan Gruber and Simon Johnson in their recent book, “Jump-Starting America: How Breakthrough Science Can Revive Economic Growth and the American Dream.” They suggest spending $100 billion a year to create a string of new research parks, along the lines of Singapore’s Biopolis, in economically declining regions with strong human resources.
Now Robert Atkinson, Mark Muro and Jacob Whiton, researchers at the Brookings Institution, have a new plan that would do something similar on a smaller scale. They would spend $100 billion over a decade to create eight to 10 regional technology clusters. These clusters would be chosen by a competitive application process. In addition to $700 million of annual research spending, each lucky city would get grants for worker training, incentives for business investment, infrastructure, tax breaks and so on. The research money would be given to local universities and distributed by the National Science Foundation.
Both of these plans have much to recommend them. Either one would raise U.S. growth, strengthen scientific and technological leadership, and help the places that received the investment. But both could be improved upon.
The MIT plan, which is 10 times bigger than the Brookings proposal, has the advantage of size. Historical rates of R&D spending suggest that the U.S. could sustain and make good use of such a large expenditure. But it has two potential weaknesses.
First, picking specific research fields for specific towns is a risky strategy. Declaring that, say, Buffalo, New York, will now be the center of synthetic biology research runs the risk that San Diego, Minneapolis or Boston is better equipped for that role. The repeated failure of governments to create rivals to Silicon Valley in the information-technology industry is a cautionary tale here.
The second problem is the focus on research parks. Evidence from the U.K. suggests that companies located in research parks earn relatively small returns, while a study of Malaysia’s parks concludes that knowledge transfer to businesses is limited. It may be that technology parks serve as a mediocre substitute for research universities in developing countries. Simply putting researchers and business people in close proximity may not accomplish much without some mechanism for producing regular interaction and informal exchange of ideas between them. Furthermore, building new technology parks from scratch runs the risk of creating hugely expensive white elephants that will be seized upon by political opponents of the spending, much as the failure of Solyndra was used to discredit government loans to clean-energy companies.
A better strategy is to simply spend research dollars through universities, as the Brookings plan would do. Evidence shows that universities revive local economies primarily through their research activities, which attract skilled workers to an area. Universities also have systems in place for facilitating tacit exchange of knowledge through seminars, networking events and casual social interactions among researchers. Most importantly, universities already exist in every region of the country; all that would be necessary would be to direct research grants at second-tier universities to bring them up to the first rank. Instead of the government picking what field of research a city would focus on, universities would be able to choose their own specializations based on existing competencies, as Carnegie Mellon University did with robotics in Pittsburgh or the University of California-San Diego did with biotechnology.
The Brookings plan has its own shortcomings, however. First, the amount of money isn't nearly enough. Second, like the MIT plan, it focuses on creating only a handful of new technology hubs. This would be good for the national economy but might fall short as a policy for spreading economic activity more equitably throughout the country. To really spread economic activity around, it needs to go to more than just a few new destinations.
Instead, new research spending should be on the order of the $100 billion a year that MIT's Gruber and Johnson propose and needs to go to a large number of existing universities in stagnant or declining regions of the country. A new government agency could also help coordinate federal research funding with state and local development efforts. A new research-based local industrial policy can succeed if it builds off of existing institutions and doesn’t put all its eggs in a few baskets.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Noah Smith is a Bloomberg Opinion columnist. He was an assistant professor of finance at Stony Brook University, and he blogs at Noahpinion.
©2019 Bloomberg L.P.