As I noted in my last post, Kahn and Glaeser found that Southerners drive more than people in other parts of the country even after controlling for metropolitan area and census tract density. I've been puzzling over that. Here are some possible explanations:
1. Southerners just like to drive more. The South birthed NASCAR, after all.
2. Southern cities tend to be less dense than cities in other parts of the country. Kahn and Glaeser supposedly controlled for that. But they assumed a linear association between density and driving -- i.e., an X% increase in MSA density leads to a Y% decrease in driving. The relationship may not be linear. Denser areas of course support better mass transit. But density surely offers increasing returns to scale to mass transit. Doubling residential density, for example, supports a denser network of bus routes, which makes buses more attractive and increases the probability that a given resident will ride the bus. Put differently, doubling residential density might increase ridership by 120% rather than just 100%.
3. Southern cities tend to be poorer, which means they have less money to spend on buses. (Of course, this also means their residents have less money to spend on cars. But of course their housing is cheaper so their residents have more money to spend on cars. And so on and so on.)
4. It's a legacy of racial discrimination. When I lived in Jackson, Mississippi, I noticed that bus riders were overwhelmingly black; I assume that whites abandoned buses once the ridership mix tipped decisively toward blacks. (The same dynamic operates in public schools). This likely has three consequences: (1) Most obviously, almost all whites drive; (2) because almost all whites drive, there is poor bus service in predominantly white neighborhoods, which reinforces whites' disinclination to ride the bus; (3) because bus service is mainly used by blacks, whites aren't inclined to vote much money for it.
I'm sure I'll come up with others five minutes after I hit the "Publish Now" button.