September 22, 2008
Housing Prices: How Far to Go until Bottom?
I'll just take the market's view here; using the futures prices from the CME (via ino.com), prices will fall about another 16% from June (or 17% in log terms):
Figure 1: Log Case-Shiller 10 city price index, (red), CME futures prices (red squares), and CPI adjusted SP Case-Shiller 10 city price index (dark blue). CPI-All rescaled to 100 2008M01-08M06. NBER defined recessions shaded gray. Dashed line indicates start of plotted futures data. Source: Standard and Poors' [xls], ino.com - real estate (accessed 21 Sep 2008), St. Louis FRED II, NBER, and author's calculations.
The latest reading from ino.com for the Case Shiller US composite futures indicates a nominal bottom in May 2010, with a reading of 151.6 (or a bottom of of 151.4 in May 2011 if you want to nitpick.
This would represent about a 40% decline (in log terms) in housing prices relative to the peak in June 2006, not too far away from the 40% (in log terms) decline I predicted in March of this year for 2009Q4, based upon the relationship between the relationship between the OFHEO and the Case-Shiller indices, and the WSJ survey of forecasts for declines in the OFHEO index up to 2009Q4.
The foregoing suggests that the market's expectations are indeed for the assets underlying the RMBS to begin rising in nominal terms around mid-2010...or mid 2011. This perspective is not new, but it's interesting to see how relatively unchanged it's been since three months ago.
Posted by Menzie Chinn at September 22, 2008 08:29 PMdigg this | reddit
I agree. Places like California could suffer with flat prices well after 2011. Prices from 1989 to 1996 were flat (7 years).
It is also interesting that the CPI adjusted 2010 price is scarcely higher than the 1990 price.
Posted by: GK at September 22, 2008 10:31 PM
I spent some time surveying apraisers in the Western U.S. in 2002-2003 on the widening housing cost/wage gap. At the time, they acknowledged widespread appraisal fraud as a common problem in their course of business. Their routine was to enter a home and ask "how much do you need?"
The house of cards created an illusion of wealth that served to mask mounting debt, and Uncle Alan promoted ATM syndrome among homeowners. Idaho may be insulated in some ways from the national projections, although we are seeing more ghost neighborhoods sprouting like weeds, particularly in destination resorts (formerly known as resort communities). Nevada is in far worse shape, but at least they have some natural immunity to gambling.
The next great opportunity for federal intervention will be to match up the displaced households I speak with on a daily basis with the holders of vacant foreclosures to negotiate short-term leases. This is a win-win alternative to bailout, houses the unhoused, and occupies the vulnerable vacant propoerties to keep adjacent values from slipping even further.
Posted by: Erik Kingston at September 23, 2008 12:47 AM
Based on my calculations of the upcoming unwind of household debt --> pushing down of GDP and the historical norm of home price-to-income, I estimate that home prices will fall 70% from year end '07 values.
Posted by: jg at September 23, 2008 07:48 AM
Minor question: Are housing futures prices an unbiased predictor of future spot prices?
Posted by: Robert Bell at September 23, 2008 07:48 AM
Robert Bell: Excellent question. As far as I know, these futures have only been in existence for about 2.5 years -- too short to do a plausible examination of bias (especially since prices have been moving one direction since then).
Posted by: Menzie Chinn at September 23, 2008 07:54 AM
Eric - I sympathize with the notion of Federal intervention to match up displaced households with distressed real estate. I would only like to suggest that this strategy has a fatal flaw: Transportation costs will rapidly escalate after 2013, when oil production no longer matches demand. Placing financially distressed families in the ex-burbs, at long distance from work with no available public transportation is a recipe for disaster. Its likely that we will subsidize this migration only to have it collapse again because it is not logistically viable.
The viability of such a plan, would require competent and dispassionate administration to match up homeowners with properties relative to future transportation and utility costs. I do not hold out much optimism that the political process in America can do so, unless war-time powers were invested in a beneficial dictator.
Posted by: MarkS at September 23, 2008 08:12 AM
Benrnake claimed this morning that the TARP reverse auction process could discover "hold-to-maturity" values for mortgage related assets. This is only the latest example of academic hubris at work. At every step, Chairman Bernanke has relied on the output of flawed models to steer the economy: from default models that told him the credit crisis would be contained to inflation models that told him the domestic output gap would contain inflation.
Now, Bernanke wants to spend $700b of the taxpayer's money in an attempt to discover the gap between "firesale" values and "true" values.
Of course, as any investor knows, a "firesale" value can reflect both illiquidity and uncertainty. Which is at work here? Well, we now that distressed debt funds today are highly liquid. In theory they should provide a non-firesale price for mortgage related assets. So, its not liquidity, and therefore it must be uncertainty that drives down the price of these securities.
How will reverse auctions diminish the uncertainty surrounding variables such as house prices, default rates. prepayment speeds, etc?
Here's what I think is going in the Chairman's mind: the effect of uncertainty must be small, because models exist that predict value with a reasonable degree of certainty, and "fire-sale" prices lie outside of the confidence interval of that model output.
In the real world, we all know now that those models are inherently flawed. The time for slavish reliance on bell-curve econometrics is past. That reliance has caused enough damage as it is.
Posted by: David Pearson at September 23, 2008 08:17 AM
What exactly are you buying and selling on this futures market? Not the homes themselves, obviously. Is it a pure prediction market, where you are simply placing bets on future prices?
jg, the only problem with your prediction is that, in many markets, the supply of homes has not kept anywhere near the demand (in terms of population and household growth). So, for example, in the New York metropolitan area, I wouldn't expect a reutrn to the historical norm of home price-to-income.
But in the exurbs of most cities, where development can and has run pretty much unimpeded, I'm with you.
And if gasoline prices are such that commuting from those exurbs is unaffordable, I'd go even farther than your prediction.
Posted by: Buzzcut at September 23, 2008 08:18 AM
Remember that, to play housing in a liquid fashion, you can always buy (and short) REITs.
Posted by: GK at September 23, 2008 09:37 AM
I'd have to agree with another 16% correction in housing prices, and yes it will definitely be different amounts throughout the various states.
$700 Billion, firstly, is not nearly enough to cover another 16% correction in housing prices, actually, it's only enough to cover another 5% drop...
The other big issue is, the $700 Billion IS NOT tax payer money. This money will have to be collected from FUTURE TAX PAYERS. This makes a huge difference. It affects inflation, currency value, economy, the financial system, housing, jobs and so on.
There are currently 50 Million homes that are vacant in the United States, and there are only 350 Million inhabitants in the USA. Supply/Demand is a major issue, as in, supply far exceeds demand, and this will also be the case in New York.
If you watch the "presentations" of Bernanke, Paulson, you can see the fear in their eyes. They know more than they are letting on, and it's NOT good news...
As you may also know, "beginning Oct. 1 for 12 months, the 1st BCT will be under the day-to-day control of U.S. Army North, the Army service component of Northern Command, as an on-call federal response force for natural or manmade emergencies and disasters, including terrorist attacks." source: armytimes.com
This doesn't mean dooms day... but it does mean get ready. If the US Financial System fails, the world as you know it will not be the same, for a lot longer than you can imagine.
The problem is, the $700 Billion "ridiculous plan" will fail if approved/passed as it isn't the right solution/fix, actually, it shouldn't even be considered in any shape, aspect or form.
Posted by: Erich Cartmen at September 23, 2008 09:45 AM
The time for slavish reliance on bell-curve econometrics is past.
"The Black Swan" is certainly at work.
Posted by: Buzzcut at September 23, 2008 09:50 AM
Who is going to buy the excess supply of houses once prices have fallen ? enough?. The marginal buyers are likely the same people who just walked away from their mortgages. Will they have credit rating sufficient to borrow again?
Posted by: Frank at September 23, 2008 12:41 PM
It actually depends on what the government does. If they continue to prop up borrowing home prices could be inflated for a long time. What would happen in this case is there would be a lot of homes empty but since the price would be propped up the builder or investor holding the house could afford to do nothing. It could be 5 to 10 years before homes find bottom. If the government allowed the market to price homes, to find their own level, it could be in 6 months or less.
Posted by: DickF at September 23, 2008 01:19 PM
The total open interest on all composite home price futures on the CMA is 32 contracts. A very cheap trade could make it say anything anyone wanted.
Posted by: Aaron at September 23, 2008 01:45 PM
Very useful information about the $700 billion and the drop in house prices it will cover. 50m vacant houses sounds incredible, though. Where's it from?
I have long opined that the tendency to postpone a downturn (from Greenspan after the dotcom bust to BB and HP in the current situation) would lead eventually to an insolvent Treasury. It has just come a lot sooner than I thought.
Posted by: don at September 23, 2008 03:38 PM
"And if gasoline prices are such that commuting from those exurbs is unaffordable, I'd go even farther than your prediction."
This is an oft repeated claim that is almost pure fallacy
The move to exurbs was a flight to AFFORDABILITY - and that has not changed ... anyone who bothers to do the simple math finds out that fuel costs are a red herring
Compare in my area - Mpls - an appx 18 mile commute to a closer in suburb with an appx 38 mile commute to an exurb ... a nice 3BR story in the closer in area is appx $350,000 (and will be an appx 30 yr old home) --- the same home in the exurb will be appx $245,000
Lets put the difference at $100,000 in price and the differnce in commute at 20 miles ea way per day - 40 mil/day RT - 200 miles per week - 800 more miles commuted per month
And lets forget that the costs for a new home are in many ways less than a 30 year old home
At 18 mpg the 800 addtl miles per week take 45 extra gals of fuel per month
At $2/gal the extra commute costs $89 a month
At $3/gal the extra commute costs $133 a month
At $4/gal the extra commute costs $178 a month
At $5/gal the extra commute costs $222 a month
The $100,000 higher mortgage to live closer in - if we use 6.5% 30 year loan - costs an extra $632 per month
By any measure it is still - even at $5 gas - dramatically cheaper to live further out
And even IF people wanted to move closer in where would they live? There is no more housing closer in. Short of tearing down neighborhoods and building condo towers the city, and suburbs, are full.
And even IF there WERE homes available - what happens when you icnrease demand witha finite supply? Simple - prices climb. Forcing flight to affordability all over.
Sorry - but simply put the whole gas cost/commute affect is simply not supported by fact
Posted by: 220mph at September 23, 2008 03:45 PM
Mortgage rates are artificailly low because of the actions of central banks, the Fed as well as Asian & petro-state central banks. Indeed, we are in a secular trough for long-term interest rates.
By 2010, this will have changed.
Posted by: algernon at September 23, 2008 07:59 PM
C'mon 220. In the northeast, anyway, the decision to live farther out often means the decision to buy a car (or another car), commute farther with both, and to drive to activities as opposed to walking, biking, or riding transit. When you add a car and change your lifestyle to include more driving, fact is that the gas cost/commute/activity effect is significant.
Posted by: firstname.lastname@example.org at September 23, 2008 08:19 PM
An earlier post claimed that there are 50 million vacant homes in the U.S.. A simple Google search would reveal that that claim is ridiculous.
There are something less than 2.5 million homes vacant, not 50 million. There are only about 111 million households!
This topic is causing some hallucinations.
Posted by: Richard Whitney at September 23, 2008 08:21 PM
Gas costs are minor in the housing price question. I believe the value of one's time spent commuting is far more significant. Calculate the difference in commute times at your rate of pay and capitalize it to see what your own time/price trade-off is.
Posted by: Zephyr at September 23, 2008 08:52 PM
Of course, the actual costs of commuting should be included in the evaluation as well.
Posted by: Zephyr at September 23, 2008 08:56 PM
"This topic is causing some hallucinations."
Indeed. People have lost all control of their emotions.
There are 2.5 million vacant homes in the US. At any given time over the last 30 years, about 1 million have been vacant as a natural state. Thus, 1.5 million are 'extra'.
The US population grows by about 1%, or about 1 million 'households' each year.
Thus, it will take 1.5 years for the number of households to increase by 1.5m and catch up with housing supply. Factor in the general caution and tainted allure of 'equity appreciation', and take that 1.5 years up to about 3 years for home demand to finally match supply. After that, new construction will cautiously resume.
Thus, 3 years from now, or 2011, is the turning point of the housing market. Some places, like California, will take much longer (5 more years, IMO). Others, like Texas and North Carolina, will be sooner.
Posted by: GK at September 24, 2008 10:43 AM
Re: 50 million vacant homes in the U.S..
The number is incorrect, and must have been a typo as everything else in that post seems accurate.
There are as of 2007, 18.6 Million Homes that are vacant, actual figures from US Government.
email@example.com the 2.5 Million Homes figure is incorrect, that figure is for the quantity of Vacant Homes currently for sale.
Ok. While posting this, I also realized, that if there are 2.6 people per household, that would come out to "virtually not realistically" 50 Million potential people living in Households to be empty.
Article from BLOOMBERG Re: 18.6 Million Vacant Homes, 2008
Posted by: 18.6 Million Vacant Homes at September 24, 2008 10:58 AM
Sorry GK, your post is completely incorrect.
The "18.6 Million Vacant Homes" posted right after you is the correct one.
Posted by: Andrew at September 24, 2008 11:01 AM
I doubt the 18.6 million vacant homes figure (even though that would benefit me, as someone who has not yet bought a home).
But if it is indeed that high, then the number of US households would have to increase 17%, from 110m to 129m, just to get supply and demand back to the general equilibrium.
It would take about 16 years for population to grow this much, so home building will not resume until 2024, by these calculations.
If it is 4.26 million homes vacant, then it will take 4 years for population to grow to meet supply, so until 2012.
So anyone who owns a home had better hope that the vacant homes are 4.26m, not 18.6m.
Posted by: GK at September 24, 2008 11:54 AM
Just so it's said:
100x monthly rent, about 3x income on the coasts... Except in the middle states, where it's already less.
That's an additional 30% down from here. The Case Shiller futures are a best case.
Don't think it can happen? Well, that's what prices used to be. If you think they've found a "new, higher plateau", you need to justify why.
The highest cost areas, like Silicon Valley, had lows of 120x and 4.5x at the lows last time. So that's your best case.
We're less than half way done.
My favorite quote about the Great Depression: "The worst thing about it was how long is lasted. At times, it seemed like it had already gone on forever, but then it'd just go on some more."
This is a LONG problem.
BTW - nice analysis on commuting costs.
Posted by: Jim D at September 24, 2008 03:56 PM
"The highest cost areas, like Silicon Valley, had lows of 120x and 4.5x at the lows last time. So that's your best case."
I think you mean 200x monthly rent. Only then does the 4.5x annual income work out at a 6% mortgage.
Do the math : $100K annual income = $450K home, which would rent for $2250/month. Not the higher $3800/month.
Posted by: GK at September 24, 2008 05:04 PM
Average pay in Silicon valley is *not* $100k. 2006 median income was $83k (source: Wikipedia), which would be a $373,500 home price, at 4.5x income. Assuming that incomes don't fall in this recession - they dropped noticeably in the last one.
As for 120x GRM, well, rent prices dropped by almost 50% during the last recession... Any floor would be set by where they eventually fall, I suppose. I'm not sure what math you're doing to determine rent prices, but they're set by market forces, and influence home prices in a down market, not the other way around. Right now, demand is outstripping supply, and prices are still rising (sharply), but job destruction is happening at a fairly fierce clip right now - HP just dropped 25,000, for instance. That's gonna leave a mark.
In much of Silicon Valley, GRM is already lower than 200x. In the more posh sections, it's still more like 400x - then again, home prices there have yet to really start dipping. I suspect that the 9% interest being charged on Jumbo loans this week may help change that.
Regardless, that's still a best case scenario. In some of them square states in the middle, 80x GRM is considered "normal". One really wonders how much lower it can go from there, since that gives profit on a rental property almost immediately.
Posted by: Jim D at September 25, 2008 06:32 AM
I'm sorry GK, you are misreading numbers.
Your article is regarding home sales, not vacant homes. Please note that vacant homes does NOT mean homes for sale.
There are indeed 18.6 Million Vacant homes in the USA as per U.S. Government Census (2008)
Website of U.S. Census:
Census Website (PDF Document)
See top of Page 3 for this exerpt.
"There were an estimated 129.4 million housing units in the United States in the first quarter 2008. Approximately 110.8 million housing units were occupied: 75.1 million by owners and 35.7 million by renters. The number of owner-occupied housing units was not significantly different than a year ago, while the number of renter-occupied units was higher than the first quarter 2007 estimate. Of the 2.1 million increase in total housing units, 1.1 million were occupied and 1.0 million were vacant units. Of the 1.0 million additional vacant units from last year, only 20.5 percent were for rent or for sale. The number of total vacant housing units, 18.6 million, was higher than the estimated number in first quarter 2007. Of these vacant housing units, 13.9 million were for year-round use and 4.7 million were for seasonal use. Approximately 4.1 million of the year-round vacant units were for rent, 2.3 million were for sale only, and the remaining 7.5 million units were vacant for a variety of other reasons."
So yes, there are 18.6 Million Vacant homes in the USA. Now if you consider that the average household houses 2.6 people as specified in an earlier post, that means there is enough housing for an additional 50 Million people in the USA.
Also the link to Bloomberg:
Bloomberg 18.6 Million Vacant Homes in USA
I hope this answers your question.
Posted by: Erich Cartmen at September 25, 2008 09:03 AM
For those who wonder what the "other" vacant homes consist of.. (copied from http://www.census.gov/hhes/www/housing/hvs/faq.html)
"Many foreclosures will be in the "vacant other" category, because they are neither for sale or for rent - they are still in the foreclosure process and tied up in legal proceedings, or being held off the market until the legal owner of the property decides what to do. In addition, it is possible the unit could be undergoing repair for future use. Also included in the "vacant other" category are units "for occasional use" and units "temporarily occupied by persons with usual residence elsewhere", both of which may contain foreclosures. Foreclosures could also be included in the seasonal category, depending on the specific situation."
Posted by: ats at September 25, 2008 01:34 PM
(continuing with the last post)
Looking more at those definitions, the 18.6 figure is meaningful to the extent it deviates from its usual average. For example, the figure was 16.3 million in 2006Q2, when all was fine.
(Source: www.census.gov/hhes/www/housing/hvs/qtr207/q207press.pdf . Subtract "occupied" from "all" .)
(It was 15.6 million in 2005Q4, when the times were even better.)
The reason the 18.6 million variable is always so large is that it includes a lot of things: 2nd homes (units for occasional use), units under repair for future use, etc., which have not necessarily gone up recently.
I think a better measure is owner vacancies.
On the whole, those numbers sort of supports GK's larger point, I think --- a surplus of about 2 million homes. But waiting out a year will not be enough for demographics to take off 1.5 million of the market, because residential construction has not gone down to zero. It still pumps in nearly a million units to the available pool of housing.
(Also, to make sure I do not mislead anyone with my previous post, not all foreclosures show up in vacancies --- it could appear anywhere, including occupied, rented, etc categories. However, the impression one gets from the census site is arguably that they often show up in "other vacant.")
Posted by: ats at September 25, 2008 02:09 PM
There are 129 million housing units in the United States, comprising owner-occupied, rented, and vacant units. Of these, 18.5 million are empty. This vacancy rate is 2.5 percentage points higher than it has been at any point in the half century the data have been tracked, translating into at least 3 million too many empty housing units in the country. This number, moreover, is rising. This is the most intractable part of the real estate bubble, for we cannot find a true bottom to home prices until this inventory of empty units starts to clear, and we cannot find a bottom to the mortgage finance market until home prices bottom out.
Each year roughly half a million homes are destroyed to make better use of the land on which they sit. Population growth also helps whittle down inventory. The household formation years--ages 25 to 34--have 39.5 million people in them forming 19 million households, a group that creates demand for 1.8 to 1.9 million units each year. On the other hand, households pass from the scene later in life, and the homes they used to live in go onto the market. There are 11.6 million households of 65- to 74-year-olds and 9 million households of 75- to 84-year-olds. Their departure increases supply by around 1.1 million units per year. On net, therefore, demographic realities add about 850,000 units to demand on top of the half-million homes that are destroyed and removed from supply.
The home building industry is in a deep recession, with additional yearly new home supply cut in half since 2006. But homebuilders are still adding nearly a million units per year. The math is simple: Build a million, tear down half a million, form 850,000 households, and the country only whittles down its excess inventory by 350,000 units per year. This is one reason to expect a further drop in new home construction, but it will still take years to get our housing inventory back to normal. The economic, social, and financial damage over that time could be staggering.
The market must deal with three simultaneous and interrelated excesses: Homebuilders made too many houses, prices rose too high, and credit standards dropped too low.
http://www.weeklystandard.com/content/public/articles/000/000/015/170jdcim.asp for full article
Posted by: paulo at September 25, 2008 06:32 PM
Don't take Mr Market's opinion at face value. The market has totally underestimated the magnitude of the housing bust so far, and I anticipate it will continue to do so.
Posted by: Professor Bear at September 26, 2008 12:31 PM
> Who is going to buy the excess supply of
> houses once prices have fallen ?
There used to be a New Deal-era rule that banks that foreclosed had to resell the house they took back within a short period of time.
The bankers fought for 40 years to get that reversed and got rid of it sometime in the around 1980, in a big omnibus bank reform law. I recall the American Banker newspaper had a front page below the fold story about teh win.
So the answer is -- nobody has to buy them.
Nation of renters.
Just like in the good old days.
Posted by: me at September 30, 2008 05:19 PM
What's the scale on the left side of the chart?
4.0 what to 5.6 what?
(I'd guess 'hundred thousand dollars" of house price but that'd just be San Francisco numbers)
Posted by: me at September 30, 2008 05:40 PM