In back-to-back Inman News columns, Bernice Ross wades into the controversy over the reporting of home price statistics. In Part 1, "Put a gag on Chicken Little," Ross argues that home prices are stabilizing, but that the media's emphasis on home price indexes published by Standard & Poor's/Case-Shiller has created "a crisis in consumer confidence." In Part 2, "Where's the beef in home-price reports? "she details the bones she has to pick with the Case-Shiller index.
Ross raises some legitimate points -- there is no doubt that some consumers don't understand that those headlines screaming about national home price declines don't necessarily reflect what's happening in their market. But in placing the blame at the feet of Case-Shiller and the media, Ross fails to get to the bottom of the question she claims to address: why do Case-Shiller's numbers differ from those produced by NAR and the regulator of Fannie Mae and Freddie Mac, OFHEO?
Before your eyes glaze over, the numbers all have their uses, their differences are easily explained, and they boil down to this:
1) Case-Shiller and OFHEO look at repeat purchases and exclude new home purchases, but OFHEO also throws in appraisals that are generated when people refinance their homes. As we have all become very cognizant of lately, what a house will appraise for and what a house will actually sell for on the market can be very different things.
2) OFHEO doesn't consider transactions involving loans that are too big or too risky to be guaranteed by Fannie and Freddie, and acknowledges that homes with these mortgages on the upper and lower price ranges are seeing bigger price declines than the homes it tracks.
3) NAR looks sales of existing homes listed by MLSs, and reports median home prices, which reduces the impact of price volatility in upper price ranges.
The result is that the Case-Shiller index can show more extreme swings in price -- both up and down -- than NAR or OFHEO's numbers.
There are actually three Case-Shiller indexes -- monthly 10- and 20-city surveys, and a quarterly report that looks at all nine U.S. Census regions. The monthly survey is the one people tend to get the most worked up about, because it looks at 20 metropolitan statistical areas that include hard-hit areas like Detroit, Las Vegas, L.A., Miami, Phoenix, San Diego and Washintgton D.C. -- all of which have experienced double-digit price declines in the last year.
To the extent that the monthly Case-Shiller index can mistakenly be interpreted -- by a cursory reading of a headline, perhaps -- as representing the state of the nation's housing markets, that's a problem. And Ross is not alone in pointing out that Case-Shiller can also miss trends in micro-markets, like Manhattan, because it doesn't track sales of condos and co-ops.
But Ross, echoing an opinon piece written by NAR Chief Economist Lawrence Yun in February, also wants to make the case that the media and Case-Shiller are purposely misleading us.
"The S&P/Case-Shiller Index is the gold (scare) standard these days for those who report on the housing market," Ross complains in her first column. "News agencies began using this index about two years ago rather than the indices provided by OFHEO (the Office of Federal Housing Enterprise Oversight) and NAR."
In his opinion piece, which was also published in the Wall Street Journal, Yun said he sees the media as being "in the business of selling news, and more sales can be made with sensationalism." Yun confided that he personally has been told "by (a) few reporters off-the-record that they are interested in increasing their viewership even if it means putting things out of context."
I'm not aware of any serious news organization that has stopped reporting numbers from NAR or OFHEO. And while some newspapers and Web sites may get a short term boost in readership by sensationalizing a story, in the long run, becoming a trusted source of information is the key to success in the news-gathering biz (television and tabloid news excepted).
Yun also goes after Robert Shiller, the Yale economist who helped develop the Case-Shiller index, claiming that as a co-founder of MacroMarkets LLC, he profits from the trade of housing and futures options on the Chicago Mercantile Exchange and has a financial incentive to "scare" the market.
"The more hedging of bets that occur, the more profits go into Dr. Shiller’s bank account," Yun claims. "And more hedging of the bets will take place if people believe there will be a crash in housing values."
What Yun neglects to mention is that price increases also spark trades, and that a reliable benchmark of housing prices is an absolute neccessity for such trading to occur. The Case-Shiller indexes are calculated by Fiserv Inc., and maintained by a committee of industry experts who follow a set of published guidelines.
Like reputable journalists, the people who publish the Case-Shiller indexes know that their product is worthless if it's perceived as biased or concocted to serve a purpose other than providing the customer with the most current, most objective information.
Ross, the author of "Waging War on Real Estate's Discounters," takes a similar poke at the credibility of the motives of those who produce the Case-Shiller indexes, claiming, NAR, OFHEO and Realogy are more trustworthy:
"Ultimately, the question is whom should you believe -- the academicians and Wall Street with their complex derivatives that gave us the subprime mess, or NAR, the federal government and the real-world numbers from publicly traded real estate companies?"
Never mind that Shiller had nothing to do with the excesses in lending during the housing boom, or that his warnings of the inevitable consequences were dismissed by the industry. Here's an excerpt from NAR's Dec. 1, 2006 Realtor Magazine:
"Robert Shiller scored instant media celebrity when his 2000 book, Irrational Exuberance, predicted the tech bubble’s explosion just weeks before the fact. Four years later, when he tried to apply the same principles to the real estate boom, he found out that all investments don’t behave alike. Shiller contended that rising home prices weren’t based in the fundamentals of population growth and supply and demand; they were bubbles, destined to pop. To the contrary, NAR economists predicted that market slowdowns would largely be gradual—a trend that’s playing out today. Shiller’s failed bubble scenario demonstrates that sometimes even smart guys get it wrong."
Yun and Ross are mostly too busy shooting the messenger to take a serious look at why the Case-Shiller numbers tend to show bigger price swings than those produced by NAR and OFHEO. Both make a big deal of the weighting formula used to draw up the Case-Shiller indexes.
Yun says the weighting "places a vastly higher weight on multimillion dollar homes" -- an approach, he claimed, that "flies directly in the face of the American sense of democratic values." That might get your blood boiling -- especially if you don't live in a multimillion dollar house -- but what does it have to do with mathematical and statistical principles that may or may not justify weighting?
According to Standard & Poor's, weighting is designed to make sure that the indexes aren't unduly influenced by "atypical" sales. A home that sells more than once in a six month period is not included in the index because it may be a fraudulent transaction, is not an arms-length deal, or follows the redevelopment of a property. Less weight is given to homes that have large changes in price relative to others in their area because the home may have been "remodeled, rebuilt or neglected."
Ross calls the system "absurd," because "Lenders don't look to computer-generated models to make lending decisions; instead they rely on those buyer and seller 'mispricing decisions' (i.e. comparable sales) to determine how much they will loan on a given property."
Hmmm... or maybe lenders should have been more careful about those buyer and seller "mispricing decisions?" Whether or not you believe it makes sense to exclude sales that might be distorting home prices because they stand out like a sore thumb from other comps, Standard & Poor's says 85 to 90 percent of sales aren't weighted at all.
OFHEO's most recent study of the differences between its home price index and the Case-Shiller indexes suggest that the effect of value-weighting high price homes is relatively small, and that the gap between the indexes is not explained by different price trends among the most expensive homes (Ross, in arguing that weighting is a problem, sights an older OFHEO study in which the authors admit that OFHEO's attempts to recreate the Case-Shiller weighting formula were "imperfect").
According to OFHEO, what seems to be propping up prices in its own index is the use of appraisal data (when borrowers refinance their homes), and also what's happening to the price of homes that are purchased with mortgages that are NOT guaranteed by Fannie and Freddie (i.e., subprime and jumbo loans). Until April, Fannie and Freddie were not buying or guaranteeing mortgages larger than $417,000. Now the government-sponsored enterprises are allowed to go up to $729,750 in high cost areas.
"Price declines seem to be particularly large for low and moderately priced homes without Enterprise-purchased mortgages," OFHEO said in its January report, "Revisiting the Differences between the OFHEO and S&P/Case-Shiller House Price Indexes: New Explanations."
Removing appraisal valuations reduced the gap between the OFHEO and Case-Shiller indexes from 4.3 to 2.7 percent. When OFHEO plugged in price data from DataQuick on sales of lower-priced homes with mortgages not backed by Fannie and Freddie, it found that the gap was closed by another 1 percent, to 1.7 percent.
So, if you are looking at purchasing a really cheap or a really expensive home in a market that saw a lot of subprime lending that's covered by the Case-Shiller index, perhaps it's not "the leader in innacuracy," as Ross claims. On the other hand, if you're looking at a home that's in the median price range, NAR's numbers might be a good bet. And if you're looking to buy an existing home in the one of the hundreds of markets covered by OFHEO that didn't see lots of crazy lending practices during the boom, that index might be a pretty good indicator of what's happening with prices.
Whichever index you choose, none paints a rosy big picture, although that doesn't stop Ross from trying.
Ross gets off on the wrong foot in her first column, beginning the discussion with an apples-to-oranges comparison between an April 29 Case-Shiller report and an April 22 OFHEO release.
The Case-Shiller price declines Ross cites (12.7 percent in the 20-city composite) are for a limited number of MSAs over a one-year period. Ross compares the Case-Shiller numbers to a slight month-to-month increase in prices (0.6 percent) for the entire country from OFHEO. Annual versus monthly. MSAs versus national. And the point is?
Ross doesn't tell us what time frame she's referring to when she states "NAR, OFHEO and Realogy all reach the same conclusion: Prices are down nationally less than 1 percent and, in many areas, prices are actually increasing."
In fact, the latest numbers from NAR show that in the last year, median home prices have fallen 8 percent (NAR doesn't report seasonally-adjusted month-to-month changes). Realogy, in its latest report to investors, says average home prices in transactions handled by Realogy Franchise Group were down 7 percent during the first quarter compared to a year ago. OFHEO puts the year-over-year average price decline at 3.1 percent, and says peak-to-trough, prices are down 3.7 percent since April, 2007.
Because OFHEO slices its monthly numbers up into nine Census divisions (rather than MSAs) and prices were up in seven of those divisions in its April 22 report, Ross claims "a whopping 77 percent of the areas in the U.S. reported a price increase between January 2008 and February 2008!"
But the same report showed year-over-year price declines in six of nine Census divisions. The latest, most detailed numbers from OFHEO (which came out after Ross wrote her columns) show first quarter price declines 43 states in the "purchase only" index (which excludes refis), and quarterly price declines in all nine Census divisions. In the "all transactions" index, which includes refis, 128 of the 292 ranked MSAs saw quarterly price declines.
