There is the "creative" city, the "consumer" city, the "mega-city", and who knows how many others. Joel Kotkin writes about the "efficient" city in today's WSJ.
While Kotkin makes international comparisons, he singles out the "big gainers" for the U.S., places with 100,000-2.5 million residents. That's a wide range and whenever "cities" are discussed, many analysts refer to metropolitan areas. I prefer the "urbanized areas" (UZAs) when looking at the U.S. case. These do not follow political boundaries, but represent more functional units. Their data are, however, only reported for the census year.
Matt Kahn comments on Kotkin's piece here. He also evokes "mellow" and "superstar" cities.
What do we know? "Cities" compete and must adapt to survive. They also specialize to varying degrees. They also include a complex mix of different spatial enclaves.
If we collect data for national aggregates, we expect that national economic growth benefits from a well-performing system of cities -- which can include many of the types listed above. And they cannot all be above average. They cannot all belong to the class of cities that performs best in some window of time. I have already commented on the amazing stability-at-the-top of city-size rankings.
As mentioned in my last post, UZA data are available for the census years, 1950-2000. No idea when we will get 2010.
For each of the decades, population size and subsequent population growth are negatively associated -- as expected. Looking at just the top 50 UZAs (in 1950), here are the five size-growth associations: 1950 size vs 50-60 growth, 1960 size vs 60-70 growth, etc.: -0.26, -0.27, -0.25, -0.14, -0.07. The effect of size on subsequent growth had been falling.
Using the 2.5 million cut-off, the sample gives us the top 15 (San Diego - New York; 2.67 to 17.8 million in 2000) and the smaller 35 (Birmingham - Minnesota; 0.66 to 2.39 million. The average 1990-2000 growth for the bigger group was 21% vs.19% for the smaller group. For that decade, bigness was not a handicap.