A recurrent theme of this blog has been to question the popular idea of "urban sprawl". No one is quite sure what it means and (far worse) it is used to spawn countless (and hopeless) statist policies (with feel-good names like "liveability" and "smart growth", etc.).
Many of us have documented the remarkable stability of commuting times. Most metro areas grow, but they find ways to accommodate the growth. Otherwise, they would not grow; capital and labor would go elsewhere. Traffic "doomsday" is (as Steve Polzin reminds us) forever impending. Land markets (imperfect though they are) must be doing something right.
Alex Anas has been doing the heavy lifting on these topics for some years via his work on computable general equilibrium models of cities. Here he applies his model to explain commuting trends. He finds that, "From the microeconomics of the model, we see how the location of jobs influences where residences are located and how, in turn, the location of residences influences where jobs locate" (p. 17).
If it were not so, we would indeed have a "sprawl" problem.