So which is it? Are home prices nationwide down by 3.1 percent or are they up by 5.9 percent? The first number comes from the National Association of Realtors, which reported that the national median existing home price fell by nearly $7,000 from the previous year, an unusually sharp decline.

The second number comes from the federal government agency that tracks home price changes in a giant housing database with over 32 million transactions, including both sales and refinancings. The Office of Federal Housing Enterprise Oversight (OFHEO) said that despite all the headlines about bubbles bursting and local housing values taking beatings, it concludes that the average home in the United States experienced 5.9 percent appreciation last year.

The agency also reported that with the exception of California, portions of the Midwest and New England, home prices generally were positive across the country. In 256 of the 282 major metropolitan markets it studied, prices rose last year; only in 25 was there net depreciation, and prices in one area remained flat.

Which study do you believe? Obviously they can't both be right. Or could they? Or maybe they're both wrong and right in part.

Here's a key clue to the puzzle: Although the two surveys sound like they're talking about value changes in the same giant housing stock, the reality is that they are measuring different houses and different things. Both surveys come with important methodological limitations as well.

NAR's widely-quoted monthly survey measures median prices by state, region, and for the nation as a whole. Its inherent limitation is its reliance on medians-mid-points, with half of all houses sold below and half above. The problem with medians is that the total pool of houses being measured can be influenced by the geographical composition of the sales in the statistical pool during a given period.

During the housing boom years, exceptionally large numbers of dwellings were selling in high-cost, high-population areas such as California and the East Coast. That inevitably pushed up the mid-point or median price nationwide. For the past two years, by contrast, sales in those former boom areas have dropped by 30 percent and more, while sales have grown in lower cost markets with big population bases, such as Texas. That inevitably is pushing the mid-point of all closed sales transactions downward.

So although it sounds grim on the TV news reports, a 3.1 percent drop in the median home price is not precisely what a lot of people may assume.

Now look at the OFHEO data. It, too, has inherent limitations. For starters, its massive database uses only Fannie Mae and Freddie Mac funded or securitized sales and refinancing transactions. That eliminates coverage of what's going on in the significant-and sometimes volatile -- “jumbo,” nonconforming segments of the national market, and under-represents places like California and the Northeast and Middle Atlantic states. At the moment, based on reports from local MLS and other real estate authorities, prices are down in many of these upper-bracket markets that boomed from 2000-2005.

The OFHEO data also omits all condominiums-again a crucial limitation because in places like Miami-South Dade, where OFHEO says average house price appreciation last year was an astounding15.3 percent-the overbuilt condo sector has seen significant price slippage because of investor panic sales and developer discounts that don't show up in the OFHEO numbers.

The NAR existing home sale data, on the other hand, does include condominiums-and thus captures those price declines.

Which numbers are right, the federal government's or NAR's? Probably the answer is that they're both correct in what they are measuring. But they happen to be measuring different things that tend to sound like they're the same when the network news covers them.

Published: March 12, 2007

Use of this article without permission is a violation of federal copyright laws.

Kenneth R. Harney writes an award-winning, nationally-syndicated column on housing and real estate from Washington, D.C. He is also managing director of the National Real Estate Development Center, a professional education company. He is a past member of the Federal Reserve Board's Consumer Advisory Council, a committee that by federal statute reviews all Fed actions on home mortgage, consmer credit and banking industry regulation. He served as a member of the U.S. Department of Housing and Urban Development's Working Group on Computerized Loan Origination (CLO) systems, and is a member of the Editorial Board of the Fannie Mae Foundation's journal, Housing Policy Debate. He is the author of two books on mortgage finance and real estate.