There is an interesting new Justin Lahart piece, here is one bit from it:
The way Princeton’s Zhenyu Gao sees it, the problem in Las Vegas was that while land was available, constraints were high enough that it took time for homes to get put up — the data on land constraints matches up well with a separate data set on how long it takes to complete the review and approval of residential projects. As a result, it took time for the supply of housing to meet demand.
But investors didn’t appreciate the fact that more supply was eventually going to arrive, and rushed in. Indeed, Mr. Gao shows that areas of the country like Las Vegas fell into a sweet spot, with more investment home purchases than both more land-constrained and less land-constrained areas. Of course, by pushing prices higher the investors invited even more development, and an eventual oversupply of homes that made the bust all the more painful.
I would put it this way: space to build plenty more, and enough slowness in the building process to get over-committed before seeing the supply-side decisions of other investors. The original research is here (pdf).