Where are home prices? Depends on who you ask, and on what they look at.
Let's say you're a homeowner in San Mateo County, California. The media tells you real estate is in the tank. You'd like to know how big a hit home prices have taken in your area.
It such a simple request.
You start with the trend in national home prices. After all, isn't the real estate market just the real estate market, one market, indivisible, with heartbreak and mortgages for all? Congratulations! You're real estate economist material! Enter grad school immediately!
Holy Toledo! Your unbelieving eyes tell you that home prices fell 8.9 percent nationally last year! You rush around the house nailing windows shut so you won't jump out. No, wait—whew!—national homes prices fell just 0.3 percent last year. No, wait, national home prices rose .8 percent last year. Sound of cautious prying. No, wait, national home prices dropped a whopping 4.6 percent last year! No, it was 5.1 percent! Furious hammering.
Desperately seeking clarity, you drill down further. Let's see, says here California home prices fell 6.6 percent. No, California home prices cratered 21.9 percent! Yikes!
Drilling down still further, you find that San Francisco Metropolitan home prices got whacked 10.8 percent last year! Bang bang bang. No, they went up 5.5 percent. Creak...creak. No, they went down 2.2 percent. Whimper. No, they went down .86 percent. No, they went down .68 percent.
Exhausted from all this hammering, prying and emotional whipsawing, you call the real estate agent who sold you your home in 2003. What quicksand did this guy get you into? Your agent tells you that the California Association of Realtors® says Bay Area home prices fell 3.9 percent last year, while DataQuick, a real estate information aggregator, says Bay Area home prices fell 8.5 percent. Still no consensus.
Closer to home, San Mateo County home prices plummeted 8.2 percent according to DataQuick.
But, your agent reminds you, real estate is local. You don't live just in the United States, or California, or the Bay Area, or even San Mateo County. No, you're a proud citizen of the upper-middle class city of Menlo Park, where the average sales price of $1,436,329 in 2007 was just 2.1 percent less than 2006's sales price and within $655 of 2005's sales price.
Well alright then. You eye your windows warily.
Furthermore, he informs you, you don't live in any old part of Menlo Park. No, you're privileged to live in an upscale part of Menlo Park called Central Menlo where, according to Multiple Listing Service data, 2007's average sales price of $2,310,041 was 2.1 percent higher than 2006's, which was 4.8 percent higher than 2005's, which was 12.5 percent higher than 2004's, which was 23.5 percent higher than 2003's. So you've done okay. You've done more than okay.
I can't tell you how relieved I am. By the way, my buddy Bob also bought in Menlo Park back in 2003, in a neighborhood he calls Belle Haven. I can't wait to tell Bob his Belle Haven home is holding steady too!
Well, no, says your agent. I sold Bob his home, and by a remarkable coincidence, Bob's home has always been worth exactly the Belle Haven average sales price, which means that his home's value dropped 6.7 percent last year, from $667,403 in 2006 to $622,444 in 2007. And frankly, it's probably worth even less today, as home prices in neighborhoods like Belle Haven continue to skid. But even Bob in Belle Haven is okay, because he paid a paltry $430,293 for his home back in 2003.
Yes, folks, it's true. Thanks to the Internet, the mass media and thousands of hard-working economists, you now know exactly home much home prices have fallen in your area. Well, not exactly. In fact, you don't have a clue. Because it depends on who you ask, and there are plenty of people to ask, and they all say different things.
And, chances are, you'll believe what you want to believe anyway. And why not, when the dismal so-called science offers you a smorgasbord of choices? Don't like your vegetables? There's always an economist who'll let you load up on the fattening stuff.
Where did the bewildering array of home price movements I gave you a few moments ago come from? From the most recent releases of the following organizations respectively:
National: the S&P/Case-Shiller® Home Price Indices; the Office of Federal Home Enterprise Oversight purchase-only Home Price Index; the OFHEO all-transactions HPI; the National Association of Realtors®; and Global Insight.
State and local: Case-Shiller; California Association of Realtors®; Case-Shiller; NAR; Radar Logic; OFHEO; and First American LoanPerformance.
In case you didn't notice, there wasn't much consensus.
How can this be? Isn't the real estate market just the real estate market? Don't white-coated economists toiling in the bowels of august institutions scientifically measure the same changes in the same market? Isn't it that simple and reassuring?
Let's look at the three most-quoted home price indices—Case-Shiller, OFHEO purchase-only and NAR—and see why they all give different answers to the same question. We'll see if they're even answering the same question.
Which index would you like to believe? Well, in the case of the National Association of Realtors, we can dismiss their numbers out of hand because we know they're lying again. That federal agency you've never heard of? It's just a bunch of bureaucrats. A Grade-5 probably scrawled their methodology on the back of an envelope back in 1933. If you're a media type or a bubblehead, you already know who you like: Case-Shiller. Because Robert Shiller was perspicacious enough to write a best-seller on the dot-com bubble, and lucky enough for it to come out the day dot-com collapsed. Which makes him an expert on real estate.
In economics, as in the real life it occasionally describes, timing is everything.
So let's look at the Case-Shiller indices first. Just out of curiosity, I'd like a show of hands: how many of you who hang on every dire utterance of the Case-Shiller indices have waded through all thirty-nine pages of their methodology? That's what I figured. So why don't we do that now?
I bet we find a few surprises.
There are twenty-three S&P/Case-Shiller Home Price Indices: twenty individual metropolitan area indices, a 20-City Composite, a 10-City Composite drawn from the twenty metros and the frequently-cited U.S. National Home Price Index which, according to S&P's latest press release, is "the leading measure of U.S. home prices". No false modesty here.
In "About the Indices", S&P tells us that they "are designed to measure the growth in value of residential real estate". The U.S. National Home Price Index "is a broad, market value-weighted composite...for the nine U.S. Census divisions". The S&P/Case-Shiller FAQs says the indices "are designed to measure increases or decreases in the market value of residential real estate in defined regions and the overall U.S. (italics mine)".
Pretty unequivocal, right? Everything you need to know about the entire U.S. housing market in a nutshell, right?
Wrong. The Case-Shiller national index "tracks the value of single-family housing within the United States". Notice something missing? That's right, the national index doesn't track condos. And according to NAR statistics, condos made up 12.6 percent of existing-home sales nationally in 2007. So that's 12.6 percent of the existing-home market we're missing when we swallow Case-Shiller hook, line and sinker.
So maybe Case-Shiller's U.S. National Home Price Index should be renamed the "U.S. National Single-family Home Price Index".
And that's not all we're missing. Case-Shiller doesn't track sales prices of new homes, which, according to the National Association of Home Builders, accounted for about 776,000 sales in 2007. Add that to NAR's estimate of existing-home sales in 2007, 5.652 million, and that's 12 percent of the total market Case-Shiller isn't counting.
So maybe Case-Shiller's U.S. National Home Price Index should be renamed the "U.S. National Existing Single-family Home Price Index".
And that's not all we're missing. Try about 30 percent of the country's existing single-family housing stock, according to S&P's own numbers. Remember its reference to the national home price index as being a composite of all nine Census divisions? Let's look first at the New England division. Case-Shiller covers only 93.5 percent of existing single-family housing values in that Census division, because it doesn't include Maine. So who'll miss a small state like Maine? But in the Middle Atlantic division, Case-Shiller tracks just 76.1 percent of existing single-family housing values. Why? Because it covers all of New Jersey but only 72.3 percent of New York and 58 percent of Pennsylvania. Case-Shiller covers only 63.3 percent of the East North Central division because it misses all of Wisconsin and Indiana and parts of three other states. The West North Central division gets just over half (53 percent) its single-family housing stock tracked, the West South Central just under half (48.7 percent). More numbers: South Atlantic 63 percent, East South Central 38.3 percent, Mountain 70.4 percent and Pacific 91.6 percent.
Speaking of the Pacific division, those of us who live in Silicon Valley will be interested to know that Case-Shiller doesn't track the San Jose metro.
So maybe Case-Shiller's U.S. National Home Price Index should be renamed the "U.S. Not-quite National Existing Single-family Home Price Index".
So alright, you say, maybe the Case-Shiller U.S. National Home Price Index doesn't track the price movements of the nation's housing stock down to the last lean-to. That'd be cost-prohibitive. Don't be so literal. Haven't you heard of sampling?
Yes, and I've also heard of sampling error. It's significant that Case-Shiller focuses on metropolitan areas, because they've experienced more price volatility than rural areas. Which would suggest that the Case-Shiller national index inherently has more volatility than the national market. Which might partially explain those eye-popping numbers that blare from the headlines and keep the bubble blogs buzzing.
And if that's not enough, Case-Shiller sampling ignores or under-represents states that are weathering the slump rather well. According to LoanPerformance's December 2007 3-Month HPI Change map of the U.S., Texas prices have edged up, yet just over half (54.3 percent) of its existing single-family housing stock makes it into Case-Shiller. Neighboring Arkansas has also done fairly well, yet has only about one in three (35.8 percent) in Case-Shiller. Indiana, which is doing surprisingly well (I thought the Rust Belt was on life support) is completely ignored by Case-Shiller. Ditto Montana, Idaho and West Virginia. New York leads the nation in home price appreciation, yet Case-Shiller counts less than three in four (72.3 percent) of its existing single-family.
On the other hand, states thoroughly hammered by the housing bust, like California and Florida, have almost all their existing single-family stock (98.2 and 97.1 percent respectively) counted by Case-Shiller. Other beat-up markets with fine representation in Case-Shiller include Nevada, Arizona, Michigan, Ohio, Massachusetts, Connecticut, New Hampshire and Maryland.
So maybe Case-Shiller's U.S. National Home Price Index should be renamed the "U.S. Not-quite National Existing Single-family Home Price Index With An Asterisk Like Roger Maris' Former Single-season Home Run Record".
Case-Shiller makes much of its repeat sales index technique, "now widely considered", it assures us, "the most accurate way to measure valuation changes". In other words, Case-Shiller doesn't compare changes in market prices en masse from one year to the next. Instead, it compares the sales price of individual single-family homes as they re-sell.
When I first heard this technique, I thought, "how could anyone find a better way to measure anomalies, the homes least typical of their markets?" Which homes are most likely to show up in a repeat-sales index? Homes bought and sold more frequently than average. Homes bought and sold by flippers, for example, who inject value. And homes bought at a premium at a market peak and then quickly unloaded on the downturn by over-extended homeowners whose distress subtracts value. What a great recipe for volatility, not the first time this word has been used in connection with these indices.
But of course, Karl E. Case and Robert J. Shiller thought of this when they "mastered" the technique as a research project back in the 1980's. Case-Shiller weights its indices to mitigate the "high turnover frequency" inherent in the repeat-sales technique with what it calls "time interval adjustments". I haven't found the place in their methodology where they share with us exactly how they do this, probably because it's proprietary. Fair enough.
But here's something even more significant: Case-Shiller also adjusts for "price anomalies". They're on the look-out for "a large change in the prices of a sales pair" (in other words, a big jump in the sales price of a home at a later sale). If they find one, "the methodology will apply smaller weights to homes that appear to have changed in quality or sales that are otherwise not representative of market price trends (italics mine)". Fine, except that how do they define a "market", and how do they decide what the correct price trend is for that market?
Is this a theoretical question? Not at all. Let's take Case-Shiller's "San Francisco" market, for example, variously described by them as either a "city" or a "region". Case-Shiller's "San Francisco" is actually the San Francisco-Oakland-Fremont Metropolitan Statistical Area which includes not only the city and county of San Francisco but also the counties of Alameda, Contra Costa, Marin and San Mateo. And you'd be hard-pressed to find a more diverse cross-section of real estate sub-markets. So if prices rocket in the posh neighborhoods of San Francisco and rise only modestly in the wide open spaces of east Contra Costa an hour away, or if prices are stable in the better parts of San Mateo County and plunging in entry-level neighborhoods just fifteen minutes away, which price trend is the anomaly within that so-called market which is, in fact, a market only for the convenience of economists who deal with large groups of numbers and not with individual buyers and sellers? Which sales pairs get their sales prices over- or under-weighted, and by how much?
And can we multiply this question by all twenty of the metros Case-Shiller covers?
But here's what I think is the piece de resistance of the Case-Shiller methodology: the statistical model assumes "that the two sale prices that make up a sales pair are imprecise, because of mispricing decisions made by homebuyers and sellers at the time of the transaction". "Mispricing?!" Yes, "mispricing variance occurs because buyers and sellers have imperfect information about the value of a property...The difficulty in assigning value to each of [hundreds of house] attributes, especially when buyers and sellers may not have complete information about each factor, means that there is significant variation in sales prices, even for homes that appear to be very similar" (italics mine). The Case-Shiller methodology uses "interval weights" to correct the "mispricing" errors made by buyers and sellers.
If anyone ever wanted to make the case that the academic economist has utter contempt for the participants of the real estate market—a contempt, I should add, shared by many who have never been participants in that market—Case-Shiller's extraordinary assertion of "mispricing" would be the keystone. According to this theory, no buyer or seller knows the correct value of the house they're buying or selling, so the academic has to wade in and set things straight. Yes, Case-Shiller seems to be saying, we'll let the marketplace think it's setting prices, since the idea has become a quaint article of faith among market-economy economists over the past two-hundred-plus years, but we know the marketplace sets the wrong prices, so we'll adjust them. Then why bother studying market prices? And are they market prices, after Case-Shiller is through "correcting" them?
If anyone ever wanted to make the case that academic economists study a wacky parallel universe of their own construction...
Let's explore the implications of Case-Shiller's "mispricing" theory for a moment. No buyer can correctly value the strength of his motivation, or the ability of his purchase to meet his needs. No buyer can know the condition of his purchase, despite property inspection reports. No buyer can know the other factors that determine his purchase's value, despite manifold seller disclosures and a general familiarity with the neighborhood—in fact, the buyer is apparently presumed to know nothing about the neighborhood he's bought in, and just stumbled across it. No buyer can know that shopping and other amenities are close by, or not close by. No buyer can know that the school he wants is just around the corner, and that it's the top-rated school in the area. No buyer can know that his purchase is convenient to work. No buyer can know that he's buying on a pleasant tree-lined street. No seller can know that she sold at market price, even though five or ten or twenty buyers competed for her home. Etc. etc. etc.
It's an attitude—the real estate consumer as perpetual dupe and doofus—so often expressed or implied in the academic's studies, yet so difficult to reconcile with the economist community's contention that the Internet has informed and empowered the real estate consumer to such an extent that he or she barely needs an agent. Trust the economists to weigh in at both extremes of this question and get it wrong both times.
This "mispricing" theory also points out one of the inherent problems the economist faces, and occasionally realizes and tries to adjust for, when attempting to measure market value. He doesn't—can't, of course—go out and look at each home in his index. And as the Case-Shiller methodology admits, the deed records it draws its information from "do not usually describe the physical characteristics of properties". The economist reminds me of the real estate agent who does a Comparative Market Analysis or Broker's Price Opinion for a home in a neighborhood in which he's never worked. He hasn't seen the homes he's comparing to the subject property, so he's not aware of the characteristics that differentiate these homes and can mean a difference of thousands or even hundreds of thousands of dollars in market value. The economist can only compare homes that "appear to be very similar", to use a phrase right out of the Case-Shiller methodology. So any inexplicable variations in the sales prices of homes that "appear to be very similar" isn't proof of the serious limitations of the economist's data-gathering and -interpreting techniques, no, it's proof that buyers and sellers don't know what they're doing.
"Economics": a game where players called "economists" get to write the rules so they always win. While their fans cheer every touchdown.
And maybe Case-Shiller's U.S. National Home Price Index should be renamed the "U.S. Not-quite National Existing Single-family Home Price Index With An Asterisk Like Roger Maris' Former Single-season Home Run Record and Based On Imprecise Market Prices (But Don't Worry We Know How To Fix That)". I wonder if the S&P Web site has a suggestions link.
I think these revelations in the methodology raise three concerns.
First, if the repeat-sales technique is so bullet-proof, why does it need so much tweaking? It's an obvious question but one I've never seen, least of all in the media outlets that routinely endorse Case-Shiller's credibility.
Second, the exact methodology for all these adjustments isn't spelled out in anything that a consumer of the Case-Shiller indices can lay his or her hands on. So it's proprietary, which is fair enough, but suddenly the process that creates the numbers is less than transparent. Numbers are being manipulated behind the scenes. All in the interests of science, but manipulated just the same. Given the average bubblehead's paranoia about the NAR's "cooked numbers", I find it interesting that no one gets squeamish when Case and Shiller play with the recipe.
And third, the more data are manipulated, the more likely that human error will rear its ugly head and the data will say things they didn't intend to. Does the Case-Shiller methodology add value or inject bias? That's not to deny that data need to be cleaned up before they're presented—as, for example, when anomalies at both ends of the range are thrown out, or when sales prices for different years or neighborhoods are adjusted to the same square footage—but even these seemingly innocuous tweaks carry risks, not the least of which is that the data can start to reflect the tweaker more than the market. Of course, the economist and his fans will assure you that he knows what he's doing. "Trust me, old boy. I've taken far more statistics classes than you."
Yeah, but I've sold far more homes than you. I have far more experience in the marketplace than you. I know the marketplace's participants far better than you. I've evaluated far more homes for far more buyers and sellers than you. Sometimes I wonder if you've ever bought a home; if you have, you seem to have gotten less out of the experience than the average twenty-something first-time buyer. I ain't buying what you're selling.
The Case-Shiller indices aren't intended to confuse or defraud, of course. S&P does make some huge claims for them, cloaked in a solemn economist-speak that somehow reminds me of the unctuous voiceovers used to peddle luxury cars. Speaking of marketing claims, I find it interesting that S&P tells us that "Case and Shiller developed the repeat sales pricing technique", while an OFHEO document we'll look at later says that "the repeat-valuations framework [was] initially proposed in the 1960s and later enhanced by Karl Case and Robert Shiller". "Developed"? "Enhanced"? What's in a word?
Certainly, most of the indices' potential drawbacks are spelled out in the methodology, which no one bothers to read and which never makes it into the S&P press releases that are passed on by the media as if they were handed down by Moses from Mount Sinai. Potential drawbacks that aren't delineated are just a matter of opinion, which is fine, this is a free country, except that no one ever points out that the Case-Shiller approach might not be the only legitimate approach. Quite the contrary, in fact.
A cynic might argue that a handful of go-getter economists have figured out that the way to get ahead is to get quoted, and that the way to get quoted is to give the media hair-raising numbers to shout from the rooftops. As reporter, editor or publisher, which factoid would you rather have for your story's lead-in: national home prices fell a spine-tingling 8.9 percent, or home prices fell a ho-hum 0.3 percent? Which claim do you think will grab more readers? Is it only coincidence that an economist who's unlocked the secret to getting on the best-seller lists says 8.9 percent and some anonymous government agency says 0.3 percent?
Then consider that the only segment of the general public that tracks these numbers with maniacal intensity is the bubbleheads, who like their housing news as gory as possible, and you realize that the economist most bearish on real estate has not only a formidable publicity machine but also a built-in fan base—although even the most rabid bubble blogger would probably laugh at the idea of making money off bubbleheads.
Speaking of the media, let's take a moment to see how they report Case-Shiller. Here's the February 26, 2008 Dow Jones MarketWatch story "Home prices post first yearly drop in 16 years", which hits us with Case-Shiller's 8.9 percent decline in the sub-head, then repeats it in the second paragraph. OFHEO's 0.3 percent decline gets second billing in the third paragraph, and I wonder how many hurried readers made it that far? Paragraph five has Shiller's grim and terse "Wherever you look, things look bleak", which tells me he hasn't looked at a number of states and local markets, including the market I work in. Paragraph six has a longer and more reasoned assessment courtesy of OFHEO's chief economist, whose name I won't mention because it's not a household name and isn't likely to become one as long as he insists on making longer and more reasoned assessments.
Twelve paragraphs into the story, MarketWatch informs us that Case-Shiller and OFHEO "are similar in construction, with a few differences". Given their dramatically different results—a 8.9 percent decline versus a 0.3 percent decline—I'd say this qualifies as laconic understatement. Then the story mentions only one of those "few differences".
Nineteen paragraphs into the story, MarketWatch tells us that the Case-Shiller "is considered by many observers to be the best gauge of national and metropolitan-area real-estate values". Funny, that sounds familiar. Oh yeah, here it is, right in S&P's "About the Indices": "the repeat sales index technique (is) now widely considered the most accurate way to measure valuation changes in various U.S. housing markets over time". Then there's the S&P February 26 press release that the MarketWatch story is based on, which says that the "S&P/Case-Shiller® Home Price Indices (are) the leading measure of U.S. home prices". And here's CNNMoney's story, also dated February 26 and also based on the S&P press release, which informs us that "the industry considers them to be among the most accurate snapshots of housing prices".
It's always gratifying to see the media fall in line behind an idea.
But isn't it odd that nowhere in S&P's publicity for the indices is there support for its claim that the whole world has pretty much decided that repeat-sales is the only way to measure changes in home prices? If everybody who's anybody is on board with this technique, then shouldn't it be not only easy but appropriate to give us a few high-profile names and endorsements? "Works for me" gushes NAR president Dick Gaylord. "Two thumbs up" grins Alan Greenspan. "Es una técnica grande y marevlous" vouches Hugo Chavez.
If I said I was "widely considered" the leading real estate agent in my area, wouldn't you want a little evidence? Not if it was convenient for you to believe that I was the leading real estate agent in my area. Not if I spoon-fed you a sensational real estate story every month. Not if I made your toes wiggle every time you tuned in to your favorite blog. Not if you were star-struck. Not if you wanted to believe with every fiber of your being.
Not every writer fawns over Case-Shiller. In fact, two of the best real estate reporters, Kenneth Harney, nationally syndicated columnist and Blanche Evans, editor of online Realty Times, have individually pointed out its drawbacks. Harney, in his December 7, 2007 "Battle of the Warring Housing Price Indices", confirms that "some of the strongest state and local housing markets today are in areas missed or incompletely covered by Case-Shiller". Evans asks "who?" when blandly assured that "many observers" swear by Case-Shiller.
Yet, oddly enough, neither have pried deeply into the methodology. Neither, for example, has mentioned or explored the implications of something Case-Shiller tells us up front: that these are value-weighted indices, which in this case means that numbers have been put through a potentially arbitrary process that attempts to compensate for the shortcomings—or "concerns", as the methodology puts it—of the repeat-sales technique. Gee, I didn't think there were any concerns/shortcomings.
Case-Shiller isn't the only home price indices that uses repeat sales. OFHEO does too, but they want you to know that there are differences. In "A Note on the Differences between the OFHEO and S&P/Case-Shiller House Price Indices", most recently published July 25, 2007, OFHEO's Andrew Leventis points out quite a few.
First of all, Leventis mentions something I noticed too: Case-Shiller is a little vague about exactly which areas its national index covers. The methodology "does not detail the specific areas for which data are not available". Blanche Evans says the data come from "100 markets" but doesn't say which markets or where she found this information. Aside from being another example of Case-Shiller's lack of crystal-clear transparency, why is this mystery noteworthy? Because "to the extent that the missing areas tend to be more rural counties, given that rural areas appear to be exhibiting strong market conditions in recent periods, the missing data might partially explain why the OFHEO and S&P/Case-Shiller national indexes diverge". In other words, "the missing data" might explain why Case-Shiller screams "catastrophe" and OFHEO mutters "eh, not so bad".
Aside from "broader geographic coverage", the other major differences between the two are that:
But here's the big difference: 8.9 percent versus 0.3 percent. Here's the bottom line: top billing versus second banana. Here's the real question: which one do you want to believe? Which one makes you shout "Yes!" inside?
Finally, let's look at the NAR's methodology, which it assures us "is really quite simple". It's also quite a bit shorter.
NAR's monthly survey of price and sales trends for the nation and for the Northeast, South, Midwest and West regions "captures 30-40% of all existing-home sales in each association/board/MLS". One-hundred-sixty associations, boards or MLSs "situated in every region of the country" participate. Raw data is "carefully evaluated by NAR economists to ensure accuracy". Examples of potentially inaccurate data are given. After screening, "the aggregated raw volume figures are weighted to accurately represent sales activity for each region". Following this is the sentence "(For more information on NAR’s weighting system)", which appears to be a non-functioning link. The weighted data are then "converted into seasonally-adjusted annualized rates".
Quarterly NAR reports for each state "uses the entire data-set (nearly 700 associations/boards/MLSs)". "Due to the significant differences in sample size for the two reports, the state estimates must be adjusted to conform with regional totals." No details are given on how the estimates are adjusted.
What about the most controversial aspect of NAR's numbers: seasonal adjustment? You may be relieved to learn that NAR doesn't seasonally adjust sales prices, although it does concede that "there is a slight degree of seasonal variation". However, "sales prices are not seasonally adjusted...[because] seasonal variances are extremely fickle and difficult to gauge". This surprises me a little—I think you could make a case that seasonal variations are relatively routine and predictable, at least in my area—but I could well be wrong, and at least we've discovered a group of economists who are sometimes willing to let the numbers be themselves.
But NAR believes it "is necessary to 'annualize' and seasonally-adjust the existing home sales data so that month-to-month and quarter-to-quarter comparisons can be observed without seasonal variances distorting the overall picture". Seasonal adjustments "are determined by using the X-11 Variant created by the Census Bureau", which should be reassuring since it makes this key part of the methodology transparent and vouched for by another well-known data-gathering organization. Why annualize the numbers? As NAR says, "home sales in the winter months are generally much slower than sales in the summer months, primarily because of differences in weather". While this is certainly true in my area, I suspect that the degree of seasonal difference varies by area, depending on the severity of its weather. For example, the San Francisco mid-Peninsula seems to run counter to most of the nation in that the peak selling period is usually spring, not summer as elsewhere, probably because our spring is so moderate (as is our winter). In fact, sales here typically slump during the summer, particularly in the midrange and top end, probably because buyers in the $1M-plus price range are more likely to remove themselves more completely from the marketplace with longer vacations spent further from home.
So which house price index is "best"? Before I try to answer that question, let's ask a far more pertinent one: why do these indexes exist?
For Case-Shiller, first let's go to the source, to a passage in S&P's FAQs that I find unintentionally amusing: "residential real estate represents a significant portion of many investors' net worth...[the indices] capture and measure this important asset class". All in the service of the goddess Science, of course. But wait: "besides tracking this critical asset class, the indices can support investors wishing to add real estate exposure without buying actual real residential property".
Peel off the academic solemnity and Case-Shiller exists for two reasons, one of which is to make a buck off speculators who want to make a buck off real estate without actually buying real estate. Next let's go to Ken Harney for a more succinct explanation. In his September 10, 2007 article "Are National Home Prices Down 3.2 Percent or Up 3.2 Percent?", Harney says that the survey "was created to provide market-by-market benchmarks for a new breed of real estate investment: Futures contracts tied to housing price movements, allowing investors to hedge their bets elsewhere on the direction of real estate values in key areas around the country".
Wall Street's been reading the bubble blogs! Because two of the themes of blog talk are: Me buy real estate? Never! Or at least not until it's the ideal time (which is blog-code for "never"); and a raging interest in speculating in financial markets. A financial market that allows real estate obsessives too paralyzed by market timing, timidity or lack of resources to speculate in real estate to, nonetheless, speculate in real estate could be a big money-maker for whoever ran that market. It remains to be seen whether Monday morning speculators are willing put their money where their mouths are.
But there's another undercurrent running through the bubble blogs that makes the bubblehead/Case-Shiller connection faintly ironic: the distrust of the "regular guy" for the real estate agent whose overweening ambition makes him say and do anything for a buck. Yet nowhere do I find among these regular guys the comprehension that the economists they've turned into gurus might also feel the pangs of ambition. In fact, ambition may be the reason an economist's profile is high enough to be recognized by regular guys.
What's the reason for the OFHEO indexes? I could quote from page 21 of its latest news release, but I like Harney's summary better: "To allow the agency, which has regulatory oversight of Fannie Mae and Freddie Mac, to keep track of the value of the collateral backing the two congressionally-chartered corporations' mortgage loan and securities portfolios".
And why does NAR publish its indexes? To deceive the public by manipulating the numbers, of course. Wait! Case-Shiller is also busily weighting and adjusting the numbers! Boy, it's hard to know who to scapegoat these days!
The politically correct answer is that "all the indexes are correct". The three indexes we've looked at differ because they measure different things in different ways.
But what it really comes down to is this: who do you trust? For the people who believe that economists are blessed with God-like omniscience, saint-like devotion to truth and monk-like indifference to the outside world, Case-Shiller is the obvious choice. Economists: not just the new witch doctors, but the new holy men.
But for those who've lost their taste for unqualified hero-worship, those who suspect that economists are merely math geeks on an unsuccessful quest for gainful employment, paragons of academic hubris whose hubris—the solemn self-assured promise of easy answers to unanswerable questions—is their only coin in the marketplace, willing yet sanctimonious traders of ideas for dollars, well, for those few skeptics the answer is also obvious.