Housing Prices: Appreciating the Difference

How much did the price of a house rise last year? Depends on where you look.

Author:: James Padinha
Publish date:: Mar 29, 1999 3:20 PM EST

Burning Down the House

JACKSON HOLE, Wyo. -- How much you reckon house prices rose last year?

The gubmint has a super-duper high-tech index that says the answer sits well below 3%.

You heard right. The statisticians at

Census

are using something called a Chain-Type Annual-Weighted Fisher Ideal Price Index in an attempt to account for the multiple measurement biases that (supposedly) plague the "common" house-price measures. Noble goal, that. Yet one does wonder: Are the bean-counters getting out much?

Check out the table above. Not a number north of 3.5% in the bunch! Does that sound realistic to members of the audience? Anyone? And check out the eyebrow-raising performances a couple of the regions show. Did the pace of appreciation really halve in the Midwest between 1997 and 1998? And did it really decelerate in the West? By almost a full percentage point?

Perhaps it's possible. But two other measures of house prices show much bigger increases and more believable paths.

Recent numbers from the

National Association of Realtors

appear in the table below. (These numbers are available at the NAR

site in the News Releases section; note that monthly sales reports contain maps as to which states belong to which regions.)

Recent numbers from

Freddie Mac

appear in the table below. (These numbers are available at the Mac

site in the News and Info section; note that the quarterly price reports contain descriptions as to which states belong to which regions.)

Surely either of these measures is more realistic than the Census series. Also, the Mac figures are probably more useful (and telling) than the NAR numbers for two reasons. One, the level of regional detail is greater (nine vs. four). The Mac numbers are even proving precise enough to pick up the fact that the desire to move to sissy cities like

Boulder

is waning (see Mountain).

And two, check out the way Mac describes its method of measurement.

Unlike other home price indexes based on mean or median values of homes sold during a given period, the Conventional Mortgage Home-Price Index is constructed, using regression techniques, from observations of actual sales prices or appraised values of the same homes over time. The street addresses of properties that serve as collateral for mortgages funded by the two secondary mortgage market firms are first processed using software certified by the United States Postal Service to create a uniform address format and are then matched to identify consecutive transactions on the same property. There are currently more than 10.2 million records in the repeat-transactions database.

Perfect? Hardly. But reasonable. And certainly more reasonable than whatever the hell the gubmint's doing.

House Party

The point here is that mortgage rates are powerful. They clocked in at a rock-bottom 6.94% for all of 1998, which goes down as their lowest annual level since at least 1972 (that's when the

Federal Home Loan Mortgage Corporation

series begins). That, alongside a solid increase in income, produced a

Housing Affordability Index

of 135.7, one of the best of the cycle. (An index of 135.7 means that the average Joe has 35.7% more income than is necessary to buy an average house at prevailing rates.) That led to a record 5.671 million-unit sales year, and that, coming on the heels of a solid sales year in 1997, produced a two-year house-price increase (gubmint stats be damned) the likes of which has not been seen since 1986-1987.

And no points for guessing what happens to spending when people see their biggest asset appreciate at a double-digit clip in two years.