“Builders slash prices to boost October sales”

The headline is gravely amiss.

Sales rose 1.7% to a seasonally adjusted annual rate of 728,000 last month from a revised 716,000 in September, which was a 11-year low.

September’s sales pace had originally been reported as 770,000. August’s sales were also revised sharply lower to 717,000, down from 735,000 estimated a month ago and from 795,000 estimated in the first release.

The numbers are being revised sharply lower each month so the presently reported month looks like a gain:

  • August – started at 795,000, revised down to 735,000 (-7.5%) in September, then revised further to 717,000 (-2.5%) as of now.
  • September – started at 770,000, and revised to 716,000 (-7.0%) today.
  • October is supposedly 728,000 right now.
  • Some simple extrapolation would suggest:

  • September’s number will be revised downward by around 14,500 (roughly -2.0%) units next month, to 701,500.
  • And two months from now, one could likely deduce that adjustments should range around -8.5%, putting actual units sold in October at around 667,000 units.

    The title will say “Builders slash prices, but home sales plummet anyway.” Eventually.

    On a side note, the Treasury Department is working on a plan to extend teaser rates on subprime loans for up to seven years – seven years is almost perpetuity! I see this as a pyrrhic victory for lax borrowers, once again concentrating on the income statement and not the balance sheet. If hundreds of billions in home loans that fueled price increases sit at low rates forever, the principal concept called “present value of money” would suggest that the true value of the loan is far smaller than the lenders booked. Either further write-downs begin, or lender balance sheets remain entirely overstated. Also note, these deals won’t extend to near-prime, HELOCs or Jumbos, which represent a far higher proportion of loans extended to borrowers during the heyday.  In addition, the valuation issue should aptly apply to the collateral for those loans as well.