The U.S. economy got a solid piece of news Tuesday morning when the Census Bureau reported a 15 percent increase in new housing starts in September relative to August, a 34.8 percent increase from a year ago. That doesn’t mean we’re in a construction boom or that one’s likely to start anytime soon. But when it comes to job-creation, rates of change matter more than levels, and even a return to a merely average level of construction could be a huge win for the economy.
But how reliable are the numbers? It turns out that estimating the total number of housing starts across the country is hard, so the data come with huge margins of error. On its face, any monthly swing in housing starts could be nothing more than statistical noise and sampling error. But there is a second set of data that are more reliable: building permits. Permits aren’t identical to starts, since some jurisdictions don’t require them and some permitted projects get delayed or canceled for one reason or another. But the two series move in the same direction—if lots of permits to start building are being issued, you can count on a lot of projects launching. And the margin of error for permitting data is much smaller and paints the same picture—permits rose 11.6 percent month-to-month and 45.1 percent on a year-to-year basis.
Housing is one of the key sectors that helps lead an economy out of the funk. That’s because home sales and home building are much more sensitive to interest rates than grocery sales or doctors’ visits, which are driven by immediate needs.
Sluggish home construction is sometimes attributed to “overbuilding” during the pre-crash boom years. Former Obama administration economist Christina Romer, for example, recently wrote that “we just built an awful lot of houses in the mid-2000s” and as a result “we are unlikely to need to do much residential construction for quite a while.” That’s become the conventional wisdom, but it doesn’t stand up to scrutiny. On a per capita basis, we’ve had at least three construction booms larger than the mid-aughts boom, but nothing even remotely comparable to the bust over the past five years.
The issue is actually slightly different. The problem is not that America has too many houses. It is that our houses are too big. Twentysomethings having trouble finding a job or suffering through low-paid internships may not want to live at home with their parents, but the physical space to accommodate them exists. And many of those children now have the financial capacity to live on their own, but only in apartments. A decade ago, young people or people with poor credit who might be doubling up with their parents were offered mortgages by a finance industry that thought it had eliminated risk. In retrospect, this was a mirage, and those homebuyers couldn’t actually afford the houses they bought. So part of the reason housing is slowly rousing itself from the doldrums is that we’re going back to building apartments. The share of new units that are in single-family structures is still higher than it was in the mid-80s or early ‘70sseventies, but it’s steadily falling. Some people own condos (I do) and others rent single family homes, but generally big structures are for renting. That trend toward multifamily construction will probably have to continue to get people out of mom’s basement and into their own rentals.
Clearly, nobody should expect residential investment to return to its 2005 peak level anytime soon. But if over the course of 2013 it merely converges to its long-term average level, that could add 1.7 percentage points of growth to next year’s GDP—great news for the economy.
That said, to many Americans a “housing recovery” means not a return of construction activity and jobs but a recovery of the housing wealth they lost in the aughts. Here the outlook is a good deal worse. The vast nationwide run-up in home prices was a genuine bubble unmoored from the fundamentals. It’s true that investors in particular locations may make money because some underlying change (better schools, new weather patterns, the rise and fall of industries) makes it more desirable. But the country as a whole has plenty of space to add new houses in response to demand for housing. In other words, the very same increase in construction activity that could boost GDP and slash unemployment should stop the typical house from increasing much in value.