In this excellent book, Harvard economist Ed Glaeser explains the economic benefits of cities. Glaeser theorizes that cities (broadly defined to include suburban areas like Silicon Valley) are more economically productive than small towns because of the advantages of proximity - that is, people learn from each other, and learn more and innovate more when they have the opportunity for face-to-face contact. He backs this up not only by pointing out that bigger cities tend to have higher GNPs, but by a common-sense argument: given that cities have higher costs, why else would companies locate in them? Some of the other questions Glaeser addresses include:
*Why do some cities thrive while others fail? He looks at New York and Detroit, suggesting that New York had two advantages. First, employment in Detroit was more centralized in a few major firms (the "Big Three" automakers) so people couldn't jump from job to job, learning new skills and forming spin-off companies. By contact, New York's financial industry is less centralized. Second, Detroit's workforce was less educated, and thus could not innovate as easily. Glaeser backs up his argument with one interesting fact: the cities with the highest number of firms per worker generally had higher employment growth in the late 20th century than cities with fewer firms per worker (i.e. more centralized economies).
*Why hasn't the Internet caused more decentralization? Glaeser suggests that high-skill jobs benefit more from face-to-face proximity. Indeed, the internet actually encourages face-to-face proximity, because people who "meet" online may eventually want to meet in person, just as people who meet by phone eventually want to meet in person.
*Why are big cities more sharply divided into rich and poor than smaller ones? On the one hand, the economic advantages of cities can create economic stars, thus increasing the number of rich. On the other hand, the huge job base of cities encourages people from rural areas and other nations to move to cities, thus increasing the number of poor. Although Third World slums may be miserable by our standards, they are still more prosperous than the villages that export people to those slums.
*How much government does a city need to prosper? Glaeser is neither a strict libertarian nor a conventional liberal. He argues that cities very much need law enforcement, transportation facilities (so their goods can reach the hinterland) and clean water. Why water? Because water-born diseases were a major killer in 19th century America, and clean water enabled people to live in cities without risking death from cholera and similar epidemics. In fact, today's cities can be healthier than their rural counterparts: Manhattanites between 25 and 34 have a 60% lower death rate than other Americans of that age, because fewer cars mean fewer car crashes, and for some reason suicides are less common as well. The gap narrows but does not completely disappear for older Americans. (Having said that, I wish that here, and elsewhere in the book, Glaeser had studied a broader range of cities and added a wider selection of statistics).
*Why are some cities more expensive than others? Glaeser suggests that land use regulation is more extensive in expensive cities, leading to fewer building permits and thus to higher prices. For example, Manhattan allowed 11,000 building permits per year between 1955 and 1964, but only 3120 per year in the 1980s and 90s. Similarly, Santa Clara County (Sillicon Valley) has high prices in part because between 2001 and 2008, it allowed only one new home for every 50 acres, while low-price Houston allowed one per five acres.
Glaeser has gotten a lot of flak for defending skyscrapers and criticizing historic preservation. But I think his critics misconstrue his argument. He isn't saying that 40-story buidings are the only way to supply affordable housing; rather, he is arguing that housing will be affordable if more of it is built somewhere in a city- whether by making more tall buildings or by making a larger number of not-so-tall buildings.