I'm completely in sympathy with Ed Glaeser's argument that a return to peak-level house prices is neither desirable nor a necessary element in macroeconomic recovery, but his take on construction is very strange:
The one sector that will not boom until housing markets come back is construction. From 2003 to 2008, the U.S. added 9.3 million units to its housing stock, and the number of vacant homes increased by 3.4 million. Construction can only come back when we work through that excess housing inventory, and that process has been slow. The rate of household formation was incredibly modest during the downturn, as young people increasingly chose to live at home. Eventually, though, construction will return to the level needed to satisfy the still growing U.S. population.
3.4 million vacant homes sounds like a lot of vacant homes. But the American population grows by about 200,000 people per month and we normally get about 0.7 houses per new person. Working through an excess inventory of 3.4 vacant homes should take about two years. And yet here we are in 2012. Now it's possible that what's happened here is that "young people increasingly chose to live at home" but a more plausible account would be that high unemployment has forced a lot of people to double up.
The paradox here is that weak employment in construction and consumer durables (appliances, furniture) is both a cause and a consequence of weak overall employment and income growth. The economy doesn't need homes to get more expensive, but it does need the Federal Reserve to clearly signal that it intends to keep interest rates nice and low for a long time to get people back to work building the new housing supply that we need to avoid scarcity and price spikes.