These guys aren’t sunny, NAR types. In the past, I thought they were overly pessimistic.
“… we find that the overall [U.S.] housing market is now undervalued 3.8 percent. That compares with a peak overvaluation of 24.5 percent during the second quarter of 2006.”
• Extreme overvaluation is now essentially nonexistent. Only three metro areas met our definition for extreme overvaluation during the third quarter of 2008, down from a peak of 52 metro areas in 2005. For the country as a whole, the housing market is slightly undervalued. When the 330 metro areas are weighted by market value, the nation is 3.8 percent undervalued. When weighted by housing units, the nation is 5.7 percent undervalued.
• Only the Pacific Northwest remains overvalued across a wide region. The protracted declines in 17 metros in California and Florida now result in their undervaluation.
Source: IHS Global Insight/National City Corp Joint Venture
Phoenix Housing Market
Click on the graphic below to see the recent history of home prices in metro Phoenix.
Phoenix house prices have gone from being 37.1% overvalued in the third quarter of 2006 to only 3.9% overvalued in the third quarter of 2008. That was a wild 2 years!
Click graphic above to enlarge.
Real Estate Geeks Only
Don’t miss “Appendix C: Metropolitan Area House Valuations / Past Price Corrections,” in the report.
This footnote is interesting, “The more severe the overvaluation, the greater the subsequent declines tend to be: correlation = +0.22.”
But this one is amazing, “The more severe the overvaluation, the shorter the duration tended to be: correlation = -0.28.” That suggests a short duration for the current housing price correction in Arizona.
{ 19 comments… read them below or add one }
Rich 12.03.08 at 12:40 pm
But this one is amazing, “The more severe the overvaluation, the shorter the duration tended to be: correlation = -0.28.” That suggests a short duration for the current housing price correction in Arizona.
No it doesn’t.
It’s a NEGATIVE correlation. That means the more severe the overvaluation, the LONGER the duration. Back to real estate geek school for you
John Wake 12.03.08 at 12:50 pm
Nope. It’s shorter.
Rich 12.03.08 at 1:10 pm
John, the fact is right in your statement:
There is a negative correlation between severe overvaluation and short duration. That means there’s a positive correlation between severe overvaluation and long duration.
I thought you just made a slip up, but now you’re showing you really don’t understand what is being said.
Sorry to be the bearer of bad news, but this report is suggesting a long downturn.
John Wake 12.03.08 at 1:27 pm
The quote from the report is “The more severe the overvaluation, the shorter the duration tended to be…”
Rich 12.03.08 at 1:48 pm
You omitted the critical part:
“The more severe the overvaluation, the shorter the duration tended to be: correlation = -0.28″
Note the minus sign, this means these 2 statements are inversely correlated. They could also have said:
“The more severe the overvaluation, the longer the duration tended to be: correlation = +0.28″.
Peter Fork 12.03.08 at 2:04 pm
Rich:
John is right. In Appendix C, the negative correlation is between overvaluation and duration, so more overvaluation means less duration.
This is not a strong correlation though, and it’s just that: a correlation, not a law.
Michael 12.03.08 at 2:39 pm
Does this include the exurbs? Because looking at Ahwatukee, it doesn’t seem too affordable. Perhaps based on 2001 and inflation. But who says 2001 was a fair value?
I moved here in 2001 and mostly lived in Texas before then. By Texas standards, real estate here seemed expensive. And all these valuation measures consider Texas to be within 10% of fair valuation.
Anyway, when you consider how dependent this area is on construction, real estate, home financing and tourism, people probably have less income (unadjusted for inflation) then they did back in 2001.
Also, Phoenix metro seems heavily dependent on small business owner compared to most places. In a recession, these folks do worse than average. In a deep recession/depression, they will be wiped out, esp. taking into account my prior paragraph.
John Wake 12.03.08 at 3:46 pm
Michael, They are looking at the whole market on average so that certainly doesn’t mean every area is affordable. And they are only looking at sales so areas like Queen Creek, Buckeye and Anthem which have very low prices now and strong sales would outweigh Ahwatukee where home sales are still dragging.
How do they determine “affordable”? Good question! This is from their report.
Our approach to determining statistically normal house values considers not only house prices and interest rates, but household incomes, population densities and any historical premiums or discounts metropolitan areas have exhibited over time.
I don’t know what’s underneath the hood but I do know they don’t lean optimistic.
Rich 12.03.08 at 4:26 pm
John,
The disagreement inspired me to actually look at the report and dig through the data. I still contend that the statement:
“The more severe the overvaluation, the shorter the duration tended to be: correlation = -0.28″
implies a negative correlation between overvaluation and short duration, but you were correct that the authors intended a negative correlation between between overvaluation and long duration.
I apologize, you were correct about their intent.
However, I’ve looked at their data and if I were an editor/referee of this paper, that statement would never have been allowed to appear as I think the data show just the opposite.
They use dozens of regional downturn data points to calculate the correlation but the data points aren’t even remotely independent. There are basically only 3 housing downturns in their data: The mid 80’s oil crash in the gulf states (possibly including WY, CO and AK), the late 80’s financial crash in the northeast and the early 90’s aerospace crash in CA. The Japanese crash in HI and the downturn in IA are the only points which don’t fall into these 3 downturns.
The only thing the correlation calculated in the paper tells us is that the oil crash was short and sharp and the aerospace crash was long and shallow.
When I isolate overvaluation by calculating the correlation within each of the three downturns, I find the correlation between overvaluation and LONG duration of +0.44 for aerospace crash, +0.4 for oil crash, +0.03 (no correlation) for the financial crash. Thus among cities under roughly similar economic climates, greater overvaluation does trend towards longer corrections.
This analysis suggests that the length of this downturn will primarily depend on external economic factors (no surprise), not the degree of overvaluation. But if we want to look at which markets will have the shortest downturns, it appears those starting with greater overvaluation will likely trend towards longer corrections.
This tells me that the underlying economic factors are FAR more
david wilson 12.03.08 at 5:19 pm
If it’s 4% undervalued now, what was it in 1999? 64% undervalued?
There’s a thing called common sense.
Use it, don’t loose it.
John Wake 12.03.08 at 5:45 pm
Like Peter said, “This is not a strong correlation though, and it’s just that: a correlation, not a law.”
Personally, the speed of this correction surprised me. I figured it would take a few more years to get to the median home price we have today and that it probably wouldn’t go that low anyway. So , I think there is some truth to their conclusion that the greater the overpricing, the faster the correction. Anyway, that’s the theory I’m operating on right now.
Time will tell.
Michael 12.03.08 at 8:07 pm
The bigger the boom, the bigger the bust…
I know there have always been far flung suburbs in Phoenix metro, but places like Maricopa, Queen Creek and the West Valley took it to a new level.
Not to mention all the people who bought houses in thoses areas were built on bubble jobs of real estate, retail and construction.
I would like to see an apples to apples affordabilty comparison based only on Phoenix, Scottsdale, Tempe, Chandler, Gilbert and Mesa.
ks 12.03.08 at 11:22 pm
The primary reason that the price correction is happening so fast in SOME places has to do with foreclosures. This is the driving force.
The driving force upward past equilibrium (long term price trend) was lax lending standards. The many loans that should not have been made caused a shortage that led to over-inflated prices. Now that the loans that should not have been made in the first place are in default, the forclusures are driving the system down with a glut of highly motivated sellers.
In the Phoenix metro we had a very fast run up. The bad loans tended to be 2 to 3 year time bombs (2 to 3 year ARMS). It should come as no surprise that a few years after the run-up, there is a “run-down”. The next leg of the run will be with the 5-year Neg. Am. time bombs. However, the government is hell bent on keeping the foreclosures at a minimum. So, who knows what the effect of the Neg. Ams will have.
Typicaly, a correlation is useless without some sort of theory as to the mechanism of correlation. For example ice cream sales correlate to crime rate. I doubt, however, that anyone would claim that ice cream has anything what so ever to do with crime.
I think that Rich is right to try and find what the correlating mechanism could be for each bubble. Talking about correlations without the mechanism is stale, boring, and somewhat useless.
John Wake 12.04.08 at 9:47 am
ks, My theory on why the greater the overpricing the shorter the correction has to do with the downward stickiness of home prices and seller capitulation.
Lesser price run-ups can take several years to resolve due to sellers slowly adjusting their prices to slower sales. When prices are greatly out of whack and sales completely tank, some sellers will capitulate.
For example, in the spring of 2007 bank-owned homes were generally priced at market. (Given the condition issues, etc. pricing at comps was overpriced for banks.) By the summer, the banks started to capitulate and price their properties much lower. If prices were only moderately overpriced in 2007, the banks may have been able to sustain pricing their properties at market price. Sales, however, were so slow at that price level in the spring of 2007, that the banks began to price their homes under market. Capitulation.
It takes a long time and a major force, but once price stickiness is broken, it really breaks free.
krogers 12.05.08 at 4:23 pm
@Rich
The author is stating the summation of results and then the actual resulting number.
You don’t correlate overvaluation and shorter duration, those are the results, not what you are comparing.
If you looked at the data in appendix C, you’d notice that the columns being compared in the raw data are valuation and correction episode (which appears to be the duration).
By correlating valuation and duration, they determined there is a negative correlation, namely -0.28, which means the greater the overvaluation the shorter the correction episode. The paper is edited correctly.
It really is not a strong correlation in this example, so as others have already said, there is a trend, it’s just not a strong one.
It only makes sense in my mind though, as what goes up must come down. It’s basic physics!
krogers 12.05.08 at 4:35 pm
@ks
According to your example, it seems you are looking more at causation than correlation.
I don’t think this paper is attempting to make such a claim in that appendix. It is merely stating that as housing prices went overvalue, they then declined, but the correction period was fairly short.
While they don’t state causation, I fail to see how housing prices do not have a direct impact on what the eventual value of the house becomes or how long it takes to get there relative to all other houses on the market.
MPS 12.06.08 at 5:15 pm
If you are looking to buy a home I think it’s a safe time to start shopping (assuming your jobs is very secure). There’s still some downside momentum left so I would be giving low ball offers if I were shopping. Neighborhoods along the light rail line should fair well in the coming years. In addition to being walking distance to the light rail my downtown phoenix neighborhood (Central Park) will be rezoned from residential to commercial mixed use in January. You can find some great investment properties here. Can you say “gentrification”?
ken44 12.13.08 at 5:08 am
Humm… seems as if the Phx metro inventory is on the rise again. Funny, because the last two years the inventory substantially dropped around this time. Any clues as to why that’s not happening this year?
MPS 12.15.08 at 6:33 pm
Ken, I’d say it is directly related to job losses and consumer fear. I think we’re in the darkest hour now.