Eric Jaffe had a good review of a University of Denver study about job sprawl. The study shows, basically, that if you make transit more convenient to people's jobs (or vice versa), they're less likely to drive to work than if you make it more convenient to their homes.
Thinking about it, that makes sense. If you own a car, you're much more likely to want to keep it at your end of the transit line than at the work end. Is there even anyone who lives in the city, keeps a car overnight in, say, North White Plains, and drives it to work in the sprawl every morning? I tried to think of actual economic or legal disincentives to this, but the real answer seems to be more primal: a car is a big expensive thing that you paid for, and you want it at the home end of the train line.
What I'm a bit less sold on is where the study authors, Gregory Kwoka, Eric Boschmann and Andrew Goetz, go from here. As Sandy Johnston highlights in a blog post, they also look at "non-work related personal trips" and conclude that working near a transit station is a bigger factor for determining whether you drive for those trips than living near a transit station is.
I can kind of see it. If you work in Manhattan, or even in downtown Denver, you're more likely to walk over to Walgreens or the Gap to pick up some necessities before you hop on the train home. You might meet friends or a date in the city after work. You might even buy groceries. On the weekend you might take the train into the city to go to a museum.
On the other hand, if you take transit out to the sprawl, you're probably just going to take transit right back rather than trying to walk around. But what I have a hard time with is the idea that when you get off the train from the sprawl and walk to your house, then you're going to get right in your car and drive to the supermarket or your kid's school or your AA meeting. I can't really see any of the New York drivers I know doing this. I know some who live in walkable communities, take the train to work in Manhattan, and then drive around on the weekends, but any car owners who live someplace walkable and work someplace not so walkable pretty much drive to work.
My guess is that this part of the study is a quirk of Denver geography. When I was last in Denver, they had only built part of the first light rail line, and there were some stations that didn't have much around them besides housing. Maybe there are a lot of stations where there isn't a supermarket or even a deli on the walk home from the train. Maybe these people live a short walk from the station but drive there anyway because there's a huge "free" park-and-ride.
I'm also not sure how this affects the land use and public investment cycles. In my experience, transit riders who don't own cars tend to be relatively strong advocates for dense development patterns as well as investment in transit and pedestrian infrastructure. People who take the train to work but drive evenings and weekends tend to identify as drivers and support sprawl zoning and investments in highways and parking lots. And sadly, even people who take transit or walk most of the time, but keep a car in the garage to drive to their country house in Vermont every few weeks, tend to identify as drivers and vote like drivers.
I think the best conclusion is that job sprawl is a slightly bigger problem than housing sprawl, but housing sprawl matters too. And even vacation sprawl matters.