US Housing

I am wondering why the rate of US house construction remains so low. I see two candidate explanations. The first is that lending standards remain very tight — the mortgage lenders having been very badly burnt once remain twice shy. The other is that potential house buyers no longer perceive owner occupied housing to be such an attractive investment and are choosing to rent or buy smaller houses.

As usual, my thoughts are stimulated by Brad DeLong who has been arguing for years that the housing bust has long since out balanced the eearly 00s boom so there are now surprisingly few single family homes in the USA

“Now the country is 4 million single-family homes short based on pre-housing-bubble trends.”

His explanation seems to be both of the above “It is clear that both these potential homeowners and the lenders are unwilling to take on the types of risk they routinely took before 2008. ”

the “both” isn’t clear to me — both parties have to be willing to take the risk for the mortgage contract to be signed. This is important, because if the issue is that potential home buyers are unwilling to make such huge risky moderate return bets, then there isn’t necessarily an efficiency loss due to financial frictions.

update: This is important and highly relevant.

I guess that the main issue is a decline in the expected appreciation of relative home prices, that is a decline in the expected return on owner occupied housing. I have two reasons — the differential between the treasury rate and the mortgage rate is normal. I would guess that increased fear among lenders would cause both tighter standards and higher differentials. Second there definitely has been a huge drop in the expected return (at least compared to the peak of the bubble when Shiller’s data start).

I don’t see a need for further explanation based on lenders’ caution.

Again, as always, I think the key issue is deciding when the housing bubble started. The standard view (which is Brad’s view) is that it started in around 2000 and that things were rational and sustainable until then. The argument is that a bubble must inflate and inflate — that a price can’t stay steadily too high because of irrational expectations about future appreciation. But in standard models even of irrational bubbles it is assumed that investors know the history of prices. Shiller claims that most people incorrectly believed that the relative price of housing had trended up in the 20th century. If he is right, home purchasing and building could have been kept high by irrational exuberance for decades. If the bubble started in the 1970s, it is not possible to estimate how many square feet of housing we have compared to pre-bubble forecasts.

Brad’s story, the generally accepted story, implies a persistent anomaly in household formation “That translates into 4 million families living in makeshift situations–primarly their relatives’ basements and attics. Yet, strangely, this enormous overhang is not exerting any pressure for a single-family housing construction recovery.”

So I decided to look up the number of households. Oddly FRED failed me — uh I mean I failed to FRED well. Google sent me here to a site which makes downloading a hassle. Hand copying, they say there were 122,459 thousand households in 2013, 116,001 in 2007 and 108,209 in 2001. So the increase is 7,792 thousand in 6 years then 6,458 in 6 years. That is about 1.4 million short not 4 million short.

I know that this calculation is ultra crude. Importantly there definitely was a pause in household formation from 2007 through 2010. However, since then, the rate has been somewhat higher than normal almost 2.5 million from 2011 to 2012 and almost 1.5 million from 2012 to 2013. I’d guess that people have moved out of basements and attics to rented apartments.

I think this might be the new normal.

I just made this figure. Note that while starts of single unit housing (that is houses) remains low, starts of units of multi unit housing has returned to normal. This fits the moving from basements to apartments story. It made the graph messy, but 5 units and over + 1 unit = about all the starts, that is starts of 2-4 unit buildings are rare (hey we own half a duplex — am I weird ?).

housingstarts