Friday, January 11, 2008
I Have a Data Point, and It's Not Pretty
by Ken Houghton
While 30% has become the standard assumption—at the AEA panel I missed most of, someone (Blinder? Roubini? UPDATE: The Bayesian Heresy notes that it was Robert Shiller) noted that a 30% drop would reduce home "values" by $6 trillion ($6,000,000,000,000)*—data is difficult to find.
Via Dr. Black, we now have a piece of data. And it doesn't exactly support the null hypothesis:
Forty cents on the dollar for land, forty cents on the dollar for homes. While I'll still grant that the plural of anecdote is not data, it remains to be seen if those points will be outliers. But the evidence doesn't favor the idea that they will be.
*This sounds as if it is a lot of money, but only partially because it is. There are two mitigating factors: (1) since it is only the appreciation, in absolute terms, over the past four to five years, it is arguably mostly froth on the bubble and (2) it assumes an overall decline in selling prices, but houses are not quick assets, so the "loss" will not be realised in any direct way, other than, for instance, a reduction in MEW.
A while back, I asked if anyone had a model that could support house prices only dropping 30%. This was after Paul Krugman called Goldman's 15% projection "improbable on the low side," and noted that a 30% drop would leave us at 2003 levels, which still strikes me as a rather high support point.
While 30% has become the standard assumption—at the AEA panel I missed most of, someone (
Via Dr. Black, we now have a piece of data. And it doesn't exactly support the null hypothesis:
Arsenault said he and his three partners may buy a block of about 50 new, unsold condominiums in Orlando, Florida. They have a price in mind and they're willing to wait until they get it: 40 cents on the dollar.
"There's a risk to buying too early in the downturn, but buying too expensive is our biggest pitfall," he said....
"They sold land at 40 cents on the dollar and they're happy to get it," Bryan said. "The value of land is eroding by the minute."...
The price, $70 million, or about $10,000 an acre, was lower than the sale price for the same land that Horton had in escrow six months ago, said Wolff Co-President Tim Wolff.
Forty cents on the dollar for land, forty cents on the dollar for homes. While I'll still grant that the plural of anecdote is not data, it remains to be seen if those points will be outliers. But the evidence doesn't favor the idea that they will be.
*This sounds as if it is a lot of money, but only partially because it is. There are two mitigating factors: (1) since it is only the appreciation, in absolute terms, over the past four to five years, it is arguably mostly froth on the bubble and (2) it assumes an overall decline in selling prices, but houses are not quick assets, so the "loss" will not be realised in any direct way, other than, for instance, a reduction in MEW.
Labels: Housing Bubble, mortgage
