November 13, 2007
New research on the causes of the housing boom and bust
What are the respective contributions of national and local factors to recent changes in house prices?
One of the striking features of home price data on which we've often commented is the fact that real estate prices during the boom rose much more dramatically in some communities than in others.
A new research paper by Federal Reserve Bank of New York researcher Marco Del Negro and University of Virginia Professor Christopher Otrok that just came out in the Journal of Monetary Economics looks at a statistical decomposition of quarterly state-level house prices over 1986-2005 into three components, corresponding to national-, regional-, and state-level factors. Although there is a single national factor in their framework, each state is allowed to respond to that common national factor with a different parameter. They find that such a framework would attribute a huge influence to the national factor over the 2001-2005 subsample. Even though prices were going up at very different rates in different communities, Del Negro and Otrok would have us think of the increases over this period as having a single, national cause.
Del Negro and Otrok then investigate what observable macroeconomic variables could account for this national factor, and find only a limited role for monetary policy.
As I've observed previously, it's awkward for me to think of this "national factor" as a self-fulfilling speculative price bubble in which each community for some arbitrary reason was responding to a national phenomenon in its own peculiar way. My interpretation is that it instead represents a national laxity of lending standards. Funds were available fairly uniformly across different communities, but stimulated different price responses in different communities owing to different local conditions for housing supply and demand.
And now the fun question is whether we will have an opportunity to update the figure above with a matching set of negative numbers.
Posted by James Hamilton at November 13, 2007 05:00 PMdigg this | reddit
Speaking of negative numbers and declining prices, John Hussman (formerly an economist) has an analytical technique for determining when we enter a recession. The technique is open book for all to see, and he shows graphs over 45+ years. Last week, the last of 4 indicators triggered, and we are entering a recession. There have been no false positives in 45 years. More info is at
Of course, we'll have to wait until June or July 2008 for the NBER to announce their analysis. In the mean time, JDH, what is your GDP growth rate analysis saying?
Posted by: Mike Laird at November 13, 2007 06:31 PM
Winston Churchill once remarked, "Never in the field of human conflict was so much owed by so many to so few."
In the instant circumstance, "never in the field of human commerce was so much owed by so many ... unnecessarily and with no ability to pay.
The legacy of one Alan Greenspan.
Posted by: esb at November 13, 2007 07:29 PM
Well, Jim, you are one of the old advocates of the idea regarding possible speculative bubbles that econometrically we can never know for sure that any apparently sharp upward movement of prices on an asset followed by an even sharper decline reflects a speculative bubble, or merely a (ahem, hack, cough) "misspecified fundamental." Now, of course there are examples that can get around this mostly, such as premia on closed-end funds that some of us have published on. But, to get really real, we do know that if we are willing to take surveys and other such evidence seriously we really can figure out when speculation is driving a market, even if it is just some local real estate market. It is when the predominantly driving agents are buying with the expectation of selling in the near future for capital gains. In the real estate markets, this manifests itself in the phenomenon of "flippers," a well-konwn phenomenon in recent years in many local real estate markets.
So, sure, lax lending standards, and very low real interest rates (which fed into various studies coming out of some of the regional Feds saying everything was just hunky-dory in the real estate markets, duh) stimulated and aggravated subprime lending and related real estate activities. Yes, not all real estate markets were speculative bubbles; large parts of the country never saw abnormal numbers of flippers (a "normal" number of which are just regular fixer-uppers most of the time). Yes, some of the local markets now in pain never had bubbles, but are just depressed regional economies, e.g. Michigan, Ohio, and so forth, suffering from chronic and worsening economic base declines with the suffering US auto industry.
But, the hard fact is that real estate markets are always local, very local. What do the real estate agents say? "Location, location, location!" So, that one cannot argue that there was some kind of coordinated national speculative real estate bubble is simply no big deal. Shiller provided good evidence in the real estate chapter in the second edition of his Irrational Exuberance of an on average national bubble. But, in the end it was local bubbles in those especially zoomy markets that was the main action, now winding down, much to the distress of many global financial entities.
Posted by: Barkley Rosser at November 13, 2007 08:22 PM
Two words: tax shield.
A renter, I think the tax benefits of 'ownership' (cough, levered long position, cough) are abysmal public policy, but I fully admit this is 2/3rds sour grapes. The other third, I assert, is golden.
Posted by: wcw at November 13, 2007 09:00 PM
Flippers, lax lending standards, bubble mentality,... all of these have to do with the demand side, and except for the labor market, most of this was national. The reason for the big differences in price changes is on the supply side. Construction costs have not gone up very much, so in places like Texas, where there is little to prevent new building, prices did not skyrocket. In places where "growth control" policies are in effect, housing supply is constricted and that allows the bubble to expand.
Posted by: Jeff Halllman at November 14, 2007 06:34 AM
It doesn't appear to me to be any correlation between the map of housing price increases and the map of forclosure rates.
The states with a lot of forclosures in the midwest are Michigan, Ohio, and Indiana. These states have had very little appreciation.
Then I look at New York, which has has a lot of appreciation (at least downstate), yet has a very low forclosure rate.
Posted by: Buzzcut at November 14, 2007 08:50 AM
Just one point. These studies that find a limited role for monetary policy in the housing bubble seem to be making an argument thereby about the lack of culpability of the Federal Reserve (Greenspan).
My point is that Greenspan's cheerleading was one factor driving the loose lending standards. He not only encouraged people to go for ARM's in the bubble, he discouraged a regulatory or even critical look at lending standards.
There are feedback loops in this situation.
Posted by: dissent at November 14, 2007 09:11 AM
Restrictions on new housing is one side of the equation. You have to have demand, specifically, population pressure on the other.
Here in Silicon Valley, the economy has been continually strong since the Dot Com bust and wasn't too bad even then. A never-ending stream of people from around the planet seek to move here for the good jobs, the delightful weather, and the social/cultural climate.
What's Texas got? Jobs, maybe.
Posted by: Joseph Somsel at November 14, 2007 10:08 AM
What's Texas got? Well, according to the Census Bureau, the number of housing units in Texas increased by 13.1 percent from 2000 to 2006. California had 7.9 percent growth over the same period. So Texas, like California, evidently had some pretty strong demand for housing. Unlike California, Texas doesn't make it difficult and expensive for builders to respond to that demand.
Posted by: Jeff Halllman at November 14, 2007 11:00 AM
It is most certainly the case that the ridiculous housing price increases in places like CA and MA, more specifically, the astronomical housing cost in "smart growth" cities like San Francisco and Boston are mainly resulted from those municipalities' strident zoning-laws that effectively and deliberately crippled the supply of housing supply. The people who suffer the most are, unsurprisingly, the lower income people, who are being forced out of the limousine liberal conclaves.
Still, I wonder what is behind the dramatic appreciation in places like NV and FL. Both states also experienced very significant housing price increase (though no way near as dramatic as CA or MA), yet both also have relatively easy going zoning laws. The price appreciation there may be mainly driven by increasing demand. Still, Arizona is experiencing the largest population growth in the U.S yet the housing price appreciation has been quite moderate there.
Posted by: aurelianoli at November 14, 2007 11:18 AM
I'll tell you what Texas doesn't have: an income tax.
Posted by: Buzzcut at November 14, 2007 12:05 PM
I would agree with the plethora of studies that implicate supply restrictions in the runup in prices in some areas, especially California and Massachusetts, where they have been documented to be very strong. Nevertheless, when you see people buying solely because they expect the price to rise soon and fast so that they can resell in the near future, one has an extra force of demand that is pushing the price up more rapidly that is purely speculative. This may have been triggered and exacerbated by the supply restriction, but the demand will rise more rapidly due to this extra speculative demand. It is well known that when the price stops rising, this extra demand tends to disappear, paving the way for sharp price declines, which we are now seeing in many of these markets.
Posted by: Barkley Rosser at November 14, 2007 12:24 PM
It seems from press accounts that the "national laxity of lending standards" itself had a strong bubble flavor. In particular, lenders and ratings agencies were using valuation models for mortgage securities, CDOs, etc that were initialized only over data during the period during which housing prices were going up. This resulted in mispricing these securities too cheaply and doing too much risky lending. This in turn led to the assets (houses) getting overpriced (since more people could borrow more money than they should have been able to, and thus bid up prices).
Thus we had a very fancy mathematical way of lending based on the assumption that prices will always continue to rise, which is JDH's definition of a bubble. It's the classic positive feedback loop - the faster prices go up, the lower the default rate (since no-one need default when their equity is increasing so fast), and the less risky lending looks, so the more leverage lenders were willing to let borrowers pile up on themselves, which in turn led prices to go up faster still.
Posted by: Stuart Staniford at November 14, 2007 01:56 PM
asset hyperinflation is worldwide..
started off by housing in US and it cascaded thru the whole world.
This was made possible by the smart (rather greedy schmucks) folks in financial world by developing many financial instruments...
As housing worldwide finally come to terms w/ this enormous bubble, financial wizards/banks will be obliterated..heck some should be forced to give back the money they earned in the last few yrs (ill gotten gains, if you ask me)
US is just the tip of the iceberg..and subprime just the tip of the entire mortgage spectrum in the US..
20 Trillion US housing can go down to 13 T...wordlwide 70 T housing can go down to 40 T..
Credit writedowns inevitably leads to DEFLATION.
world enjoyed debt orgy for a long time...welcome to severe hangover...
financial engineering/banks have cause this cancer in the world..They will be made to pay for it.
Posted by: andiron at November 15, 2007 04:50 AM
"Smart Growth" is being tarred unfairly by you (and many other right-wingers); it's typically an attempt to loosen zoning codes, not tighten them (but loosening them in the sense of allowing more density in core areas, not just allowing endless sprawl). Needless to say, neither San Francisco nor Boston have engaged in anything like "smart growth". Try Austin, Portland, etc.
Posted by: M1EK at November 15, 2007 06:25 AM
I thought that "smart growth" was the same thing as "growth control". Thanks for the clarification.
Posted by: aurelianoli at November 15, 2007 06:52 AM
a-nol and others:
To reiterate m1ek's comments on smart growth: It is NOT a tightening of zoning codes. It is an attempt to give flexibility to zoning codes or replace them altogether with performance-based codes to enable a more market-based (i.e. less restrictive regulation) approach to development. Massachusetts has very few smart growth laws. Local zoning did have an effect on prices until interest rates went insanely low and banks started giving credit to anyone who asked. The effect on prices was the same in MA as it was in the zoning-free NV desert.
Posted by: waiting_ at November 15, 2007 07:42 AM
it's typically an attempt to loosen zoning codes, not tighten them (but loosening them in the sense of allowing more density in core areas, not just allowing endless sprawl).
Well, that's what it should be, and I wholeheartedly agree that zoning should be loosened to increase density in urban centers. There's a lot of zoning policies designed to preserve the housing of the extremely wealthy in urban cores and keep out high density condos and apartments that would be good for mass transportation and the environment.
However, Mr. Dahmus, the mantra of "smart growth" has definitely been adopted by people who simply wish to restrict growth, so the tarring is not completely unfair. I've definitely seen the term used by those who are actively fighting against higher density in urban cores. For them "smart growth" means growth managed so as to not let density get too high. Now, we believe that that's a disaster and a completely reversal of the idea, but it also goes to show the problem with adopting a nice-sounding euphemism like "smart growth" to mean a particular policy. "Smart growth" can mean a lot of different things to a lot of people; some people think "smart" means preserving the historic character of a low density wealthy urban neighborhood like Georgetown, or Manhattan brownstones.
Although there is a single national factor in their framework, each state is allowed to respond to that common national factor with a different parameter.
The only problem with this approach is that, as they admit, "[s]ome states, like Iowa, Nebraska, or Oklahoma, are barely affected by the common cycle, while others, for instance most states in the North-east, are strongly affected." The question then is whether the set of states strongly affected by the common cycle in fact form another region with shared regional effects, perhaps from having similar industries. The only groupings of states that they use are regional ones, but of course it is possible that there are regions of similar states not in geographically compact regions.
A national effect that strongly affects one group of states and hardly affects at all others is inherently somewhat suspect, particularly when the strength of the "national effect" is much smaller outside the last five years, though obviously there's still something important going on if there's a regional factor that affects a "region" of twenty-odd states of so in the last five years.
The assumption of independence of the various factors is also questionable, except of course that it makes the mathematics much more possible to do.
The simpler way to put most of this is that there's still a question of correlation versus causation. They've demonstrated that the group of states that all had rapid asset increases in the five year bubble period all had rapid asset increases in the five year bubble period, and that the states without such increases were not that much affected by that national factor, and that that national factor was not that significant outside the bubble period. It would still be possible for that group of price appreciating states to all have local factors around the same time, which would produce the national factor mentioned. It seems more likely to me that there is a national factor, or some other common factor affecting the large group of states, that appeared during those years.
Posted by: John Thacker at November 15, 2007 08:09 AM
Of course, it would be equally cherry-picking to simply pick all the states for which \beta^0 is large and call them a region, but I think it's worth trying to find out if those states have interlinked economic factors or something else that would cause them to move in concert, unlike the states with a low \beta^0.
Posted by: John Thacker at November 15, 2007 08:12 AM
I would like to dig a little deeper into the appreciation / foreclosure relationship.
Perhaps prices appreciated more rapidly in areas where the population had more disposable income to invest in housing. Certainly that would hold true in places like MA, NY, CA and even MN, all areas with high levels of appreciation.
Correspondingly foreclosure rates have been high in communities where appreciation was mild, again due to lack of disposable income.
In short prices rose where people had the money to afford it and they continue to be able to make payments despite rising costs. In places where the population at large was already strapped borrowers were not upsizing. In stead they were refinancing to afford their smaller properties. Now rising rates are punishing those borrowers.
Anything to this?
Posted by: JS at November 15, 2007 08:57 AM
“. . . so in places like Texas, where there is little to prevent new building, prices did not skyrocket. In places where "growth control" policies are in effect, housing supply is constricted and that allows the bubble to expand.”
You're argument sounds plausible, until you examine other places with little or no "growth control", such as Bakersfield, which experienced rapid price appreciation.
(Note: Bakersfield is just a symbol. As noted above, look at Las Vegas, Florida, Arizona, etc. for other examples).
Posted by: MTHood at November 15, 2007 09:28 AM
No need to look with hindsight at for reasons for home price arreciation in the US. Just look at what's going on in Toronro right now.
People lined up around the block for a week? On behalf of real estate agents? A maximum of 3 units for each agent?
No. No speculation here.
Posted by: Insurance Guy at November 15, 2007 09:38 AM
Similarly, RE/Max has acknowledged 50% of condos in Vancouver are being bought by speculators.
I'm looking to Fla. to see how that plays out over time.
Posted by: Benny Lava at November 15, 2007 11:56 AM
I think we can agree that there are multiple factors involved in the recent appreciation in real estate prices.
Some are local (zoning, specific employment markets creating disposable income or not), some are regional (migration, industrial changes), some are national (interest rates) and some are cultural (speculative frenzy; see Toronto and other countries.) Others cross boundaries (tax treatments.)
My original point was that it is difficult for housing to appreciate in the face of population out-migration - please don't ignore that one. I'll grant that Texas doesn't have that problem and that Texans pay lower taxes than Californians but that makes housing debt as a tax shelter less appealing.
I hope we agree that the important point is that Congressional action and interference is dangerous. Please let the market make the adjustments necessary.
The Market knows best!
Posted by: Joseph Somsel at November 16, 2007 10:29 AM
I've seen three different reports over the last few days that shed a lot of light on this - one from Fitch, one from S&P, and one from the CEO of Wells Fargo. The picture they paint is that quality of originations was the most important thing (preventing appraisal fraud, accurate income assessment, etc), CLTV affects loss severity a lot and has much less affect on loss frequency for owner occupied (I imagine that changes a lot with non-owner occupied), vintage matters a lot, ADR resets matter a lot, and FICO was overrated. Below is my attempt to post links (some may require free registration, navigating through the web sites).
Drivers of 2006 subprime vintage performance
S&P evaluates the Levels 6.1 Model
Well Fargo CEO to investors
Posted by: Josh Stern at November 16, 2007 10:55 AM
Lending standards were certainly part of the problem. But I think you have to dig deeper.
Why were investors so willing to buy MBS and ABS - to the point that nobody cared what was going on with actual originators? Why were borrowers so eager to buy clearly over-valued homes - to the point of committing fraud in the mortgage application process on a pretty braod scale though stated income and no doc loans?
Posted by: Insurance Guy at November 16, 2007 11:57 AM
Yes, despite the context, I didn't intend my comment above to be an explanation of residential pricing patterns since 2001, or whatever was considered to be the start of the boom. Rather focusing on the nature of the lending mistakes plays a special role in explaining the nature of the last incremental deals that shouldn't have closed, and where the subsequent credit tightening was required. Also, I've read so must on the credit crunch and writedown problems that I tend to think of that as the big boom/bust phenomena, equally significant and related to housing.
As part of explaining the last blow off top and subsequent downturn in housing, 'marginal' lending (in all three senses of the word) competes with and supplements factors like "increasing mortgage rates and their lagging effect", "asset bubble market psychology", "overbuilding", "consumer stresses due to higher prices", and "lagging effects of higher property taxes".
An interesting question is why investor demand for ABS should increase at a time of rising interest rates. I haven't really come across a convincing story about that. I wonder if the rise of internet banking might have played some role in causing people to have more assets in money markets due to the increased convenience of moving funds in and out of them on the internet?
Posted by: Josh Stern at November 16, 2007 12:35 PM
Bakersfield is not a very good counter example. Its true that there was a lot of construction there and that prices went up at the same time. But I suspect that the new homes were larger and more luxurious than the existing ones, and thus some of the price increases were an illusion. And while Bakersfield didn't obstruct development, so many other California localities did that massive numbers of businesses and residents ended up relocating to Bakersfield. In effect, you had one locality absorbing pent-up growth from all over the state, and even the massive amount of new construction couldn't keep pace until early 2006. It's worth noting that the price declines started earlier in Bakersfield than in the rest of the state, and that the absolute level of prices there never got to the stratosperic numbers found in coastal areas.
Posted by: Jeff Hallman at November 16, 2007 05:10 PM
The thing never mentioned in all of this was the fraud committed by many. Some folks bought multiple homes. Fraud rings committed 10's of millions in fraud in many of these hot regions. Mortgage Brokers easily set up shop and hired "boiler room" types to move the most profitable (and generally worst products) to uninformed consumers.
Posted by: Oscar Toscano at November 16, 2007 11:01 PM
Fraud and dishonesty occurred all through the chain from borrower to investor. Borrowers lied on a fairly grand scale when reporting income for stated income and no doc loans, mortgage brokers steered borrowers into higher margin loans and amended loan applications to ensure approval, investment banks turned a blind eye to inadequate underwriting procedures and thereby failed their due diligence obligation to investors, and there is some question as to whether the rating agencies were influenced by conflicts of interest.
I found a post dated from back in March by Professor Hamilton in response to a discussion as to what role speculation had played in the boom (copied below). It seems to me that Professor Hamilton accepts that some price speculation contributed to the general price appreciation, but that market failure contributed to the problem by setting expectations that were unrealistic. I think that's a very interesting observation that I had not understood in his argument previously. I would note though, that fraud and dishonesty accompany every speculative "bubble", so its hard to say which came first, the chicken or the egg.
Professor - if your reading this, maybe I haven't understood/explained your point exactly. If you would like to comment, I would be very much obliged.
"mort_fin, I would include adaptive expectations models (in which the expected price appreciation is a function of past price appreciation) as an economic fundamentals model, since, when you solve it out, the price is a function of the current and past fundamentals. A rational-expectations and adaptive-expectations model could behave similarly or could behave quite differently, depending on the parameters. However, when they behave very differently, one then gets into the further question of how the lenders of mortgages could base their expectations on something so far removed from what happens. And then I think you are getting back to the issue of possible market failure, which is where I'm trying to steer the policy discussion"
Posted by: Anonymous at November 17, 2007 05:58 AM
That last post was me.
Posted by: Insurance Guy at November 17, 2007 05:59 AM