As incredible and out-of-balance as the housing bubble was, I'm becoming more convinced that the housing bust -- particularly in new home construction -- has been even more remarkable.
From 1960 to 2007, America built an average of about 1.5 million homes per year. That ballooned to 2.2 million per year during the housing bubble, and collapsed to half a million a year in 2010. After rebounding from the depths, about 800,000 new homes were built in 2012.
There's something incredible about that 800,000 figure. I recently came across this chart in a paper by UCLA economist Ed Learner, showing housing starts from 1920-1950. You'll notice that housing starts in the year 1925 were about ... 800,000.
Think about that. We're building the same number of homes today as we were back when cars had to be started with a hand crank, and the new technological breakthrough was a giant box called a "radio." The U.S. population in 1925 was 115 million and growing by about 1.5 million per year. Today it's 315 million and growing by 2.5 million per year.
Comparing current housing starts to 1925 levels isn't meaningful in itself. It's just a way to put the current market in perspective. But that perspective should be powerful: Given household formation and current levels of existing inventory, there is no reasonable way you can justify current levels of housing construction. Barring a deep demographic shock like a war, it won't support population growth.
Bill Miller of Legg Mason told the Financial Times this month:
He [Miller] says there is a big structural demand for homes due to a growing population and a lack of building during the bust, when fewer than 500,000 new homes a year were built. The long-term trend is for 1.4m to 1.5m new homes a year, so to catch up "we probably need to get to 2m housing starts at some point in the next five years."
We don't know exactly when that's going to happen. It could be next year, or five years from now -- maybe even longer. But we know with a high degree of confidence that it will happen someday. People need a roof over their heads.
And we know what will come with a boom in construction: It's good for jobs, it's good for economic growth, and it's obviously good for homebuilders. A quality builder like NVR (NVR 0.77%) has a good chance of outperforming in the coming years.
Some worry that homebuilder stocks look expensive. That's understandable. But valuing a homebuilder based on current earnings might be misleading. If housing starts double from current levels, as Miller and others suggest, earnings growth at homebuilders will surge. A decade ago, people made the opposite mistake, assuming homebuilders were cheap based on inflated earnings. Remember: Busts can be just as distorting as booms.
