Google, turn your gigabyte city into a science experiment
Chattanooga, Tennessee is running an interesting experiment . By the end of this year, its residents and firms will be able to opt into accessing an extremely high speed (1 gig per second) Internet access. This is a specific place based treatment. Will this stimulate economic growth or simply stimulate increased downloads of movies, videos, and images of Lady Gaga and other leisure activities that raise well being but won't "create jobs".
Cities have tried out many placed based strategies for encouraging growth. Favorites include; new sports stadiums, new arts museums and Ed Glaeser's favorite --- building billion dollar federally subsidized rail transit systems). Could fast Internet cable achieve what these other strategies have not? I have blogged about this subject before .
But, now a new point. Google is undertaking a similar effort.
"Eric E. Schmidt, Google’s chief executive, and other leaders in technology and government point to the trailing broadband performance as a danger to American competitiveness that threatens to saddle the nation with an “innovation deficit” compared with other countries.
To help close the gap, Google pledged this year to supply service at one gigabit a second to up to 500,000 people in the United States. The company says that 1,100 communities have applied, and Google will make its selection — one community, or a few — this year."
PERMIT me to make an interesting point. Suppose that Google chooses to place this "gigabit Internet" in 50 small cities. For each of these cities, Google will have identified a second place "runner up" who just missed being treated with this fast Internet cable.
A comparison of subsequent economic growth and quality of life in the 50 "treated" cities versus the 50 runner-up "control" cities --- would yield an estimate of the average treatment effect due to fast Internet Access. Note the sharp regression discontinuity here. This is no ugly "propensity score" design. This empirical strategy exactly matches the approach adopted by Michael Greenstone and Enrico Moretti in their important paper on the causal effects for rural towns when they lure a Manufacturing plant to locate there.
In the name of Reseach, I ask my friends at Google -- do not choose one site for your fast Internet. Yes, you save money paying only one fixed cost of delivering it there. But if only 1 city is treated, then there is no way to study the causal impact of your treatment.
For small and medium sized cities, does fast Internet access offer spillover benefits as an endogenous cluster emerges and co-agglomerates? For example, will movie makers and software designers and good restaurants emerge? To answer this, we need some serious thought about co-aggomeration. Examples of recent relevant urban research using manufacturing include;
Ellison et. al
Greenstone, Moretti and Hornbeck
but we need some similar work on service clusters. See Jed Kolko's work on this subject .
------------------------------
The Christian Science Monitor has assembled a diverse group of the best economy-related bloggers out there. Our guest bloggers are not employed or directed by the Monitor and the views expressed are the bloggers' own, as is responsibility for the content of their blogs. To contact us about a blogger, click here.
