Missing Something Important

I've seen a few posts on how housing starts have dropped, but since population is still growing this is unsustainable and will undergo a serious correction. Matt Yglesias follows up here with a surprisingly bad (though not, I suppose for MattY haters) graph. People per new housing start is a really bad ratio. If the population suddenly stopped growing, then we would need no new housing starts and so that would stop and that ratio would skyrocket, but would be completely and totally sustainable.

The appropriate graph, and one I haven't seen anywhere, would be population to total housing (including condos, apartments) ratio. That ratio should be more or less constant. Increasing wealth will shift it to more housing per person as there would be less sharing and more second/vacation home ownership, but as more wealth also often means slower population growth it could end up shifting both sides of the fraction. Still, that is the relevant graph.

And so long as that ratio stays above trend there is no reason for new housing starts to pick up in any meaningful way. Moreover, so long as the pickup leads--even a little--the return to trend, there is no reason to expect a major housing correction.

Lastly, of course, even though housing has national trending and inside the US mobility makes for some balance, housing is a local phenomenon, so some places could pick up while others stagnate for a century.