How Many Charter Cities Can Succeed?

+ Paul Romer

A friend wrote in to ask, “Can more than a handful of charter cities succeed?” The best way to answer his question is to pose a slightly different one. “How many big cities can there be?”

The first question implicitly presumes a winner-take-all competition where only a few cities survive. The increasing returns associated with successful cities tend to encourage this kind of thinking. Big cities seem to have permanent advantages and some types of economic activity concentrate in just a few places. For example, there may only be a handful of major financial capitals like London, New York, Tokyo, and Shanghai.

But this way of thinking misses three basic facts: the world’s population will reach at least 9 billion people, most people will end up living in cities, and cities don’t need to become financial hubs in order to grow large and succeed. The facts suggest the world can potentially support a large number of vibrant, prosperous cities.

Here’s a simple way to think about the possibilities. According to the U.S. Bureau of the Census, in 2008, of the roughly 300 million people living in the United States, about 75 million, or one quarter, live in the 9 U.S. cities with populations in excess of 5 million:

New York, 19 million

Los Angles, 13 million

Chicago, 10 million

Dallas, 6 million

Philadelphia, 6 million

Houston, 6 million

Miami, 5 million

Atlanta, 5 million

Washington, 5 million

Suppose the world’s patterns of urbanization ultimately resemble those of the U.S. With 9 billion people, the world population will be 30 times the current U.S. population. This suggests that about one quarter of the world population, or 2.25 billion people, will live in the 30×9=270 world cities with populations exceeding 5 million.

If well-run cities started to compete for residents from developing countries, how many people could they attract? A couple of billion. How many of these cities could grow large and thrive? Hundreds.

Photo by: Arenamontanus
Photo by: Arenamontanus
Back to top
see comments ()