Kevin Erdmann has produced a report for the Mercatus Center, discussing the housing bubble. Here is an excerpt:
Contrary to Chairman Bernanke’s assumption, at the national level there was no overhang of housing supply that needed to be worked off in 2011. Indeed, even in 2005 there was no national oversupply of housing. Rather, the American economy was burdened by a shortage of housing, especially in the Closed Access cities.
The housing bubble was concentrated in cities in the coastal Northeast, California, Nevada, Arizona, and Florida. Limiting our analysis to the 20 largest metropolitan areas, the Closed Access cities make up three-quarters of the “bubble” cities, in terms of total real estate valuation. Constrained housing supply was clearly the primary source of high prices in those cities, not excess demand. Prices in the Closed Access cities today remain as high relative to other cities as they were during the bubble because constrained supply is the fundamental reason for those high prices, not reckless credit markets.7
Even in other bubble cities with generous building policies, the primary cause of rising prices was the severe Closed Access shortage of housing. This is because those other bubble cities were the main destinations for households migrating out of the Closed Access cities. I call those cities Contagion cities, because in spite of their more generous building policies, they were overwhelmed by the problem created by the Closed Access cities. In the years leading up to the financial crisis, the shortage of housing in the Closed Access cities had become so severe that each year hundreds of thousands of households moved away in search of an affordable home. Many of them landed in inland California, Nevada, Arizona, and Florida.8
Figure 7 compares net domestic migration of Closed Access cities and Contagion cities. Notice that high rates of out-migration from Closed Access cities correspond to periods of large in-migration to the Contagion cities. Credit markets may have facilitated some of the housing activity during the housing bubble, but at its core this was a mass migration event caused by a lack of housing.
When Kevin first started working on his housing project, home prices in America were still severely depressed. At that time, many found his claims about the “bubble” to be rather far-fetched. Since then, home prices have recovered in many key cities and there is increasing agreement that supply constraints, not excessive demand, are the fundamental problem.
Kevin has also produced a comment on federal housing regulation for Mercatus. In this piece Kevin is more supportive of the GSEs than I am. One thing I like about Mercatus is that there is no official party line; we have a diversity of viewpoints.
READER COMMENTS
Alan Goldhammer
Jun 6 2018 at 8:37am
Scott writes,
Given that the largest funder of the Center has come out against tariffs, one wonders if there will be a diversity of opinions on that issue.
Tom DeMeo
Jun 6 2018 at 10:00am
These are places that have evolved over multiple generations, and they are generally acknowledged as so successful that they drive innovation in the broader culture.
Should a municipality seek a price equilibrium, and metastasize until it becomes sufficiently unpleasant that it stops growing? Or should it seek a size that optimizes its utility to its existing citizens, and remain a stable place? Is it really reasonable to point to a place and say it is closed because it doesn’t endlessly seek to become more densely populated, forcing constant change at a disruptive pace?
robc
Jun 6 2018 at 2:22pm
Does the place own all the land or are they preventing other owners from changing as they wish?
If they former, then no, if the latter, hell yes.
“distruptive pace” sounds like schumpeterian change. Sounds like a good thing to me.
Richard
Jun 6 2018 at 2:42pm
If there was no bubble, and supply constraints drove up prices, then why did house prices fall precipitously from 2006 to 2008, by most measures not finding a bottom until 2010 or 2011? Were supply constraints suddenly relaxed then? Surely not. Housing starts declined during that period, indicating that prices fell due to reduced demand, not increased supply.
Kevin Erdmann
Jun 6 2018 at 4:07pm
Richard: Exactly. We didn’t have a credit bubble followed by a supply induced crash. We had a supply shortage followed by a negative credit shock.
Kevin Erdmann
Jun 6 2018 at 4:19pm
Tom: That’s an excellent question. Two answers are:
1) The reason this tension exists is because density is making these urban cores more valuable, not less.
2) The high value of the urban properties reflects monopoly rents. It isn’t a reflection of value. It is a reflection of a transfer from consumer surplus to rentier profit.
Your challenge is an important hypothetical, and in a different context where there would be much more building and density, with market values much closer to the unencumbered cost of building, there might be a point where the value of existing properties is being unduly reduced by adding new properties.
In the current context, we have a collective action problem where the city would be better off with growth but individual owners can block that growth for their idiosyncratic benefit. More building would raise the value of the real estate in the metro area while at the same time it would reduce monopoly rents to existing owners.
In fact, there are many people who argue that the agglomeration benefits of density are so great that adding more dense housing will increase the market value of existing properties. (Maddeningly, sometimes this is used as a reason not to build, in order to keep housing “affordable”.) This argument, in all its forms, is wrong. Clearly there is a normal relationship between supply and cost that would drive prices down with new supply. But, they are on to something. There is some gain from density, but in adding supply and increasing density, gains in value would be mitigated by reduced monopoly rents.
Kevin Erdmann
Jun 6 2018 at 4:40pm
Scott,
Agreed, regarding Mercatus. I am thoroughly impressed and humbled by their (and your) support for helping me refine disciplined conclusions from my work, even when it is contrary to other points of view that have been expressed in other Mercatus publications.
It also highlights something about some commentators and authors who paint Mercatus as some sort of propaganda machine. The difference between their perception and reality, as far as I can tell, is like night and day.
Richard
Jun 6 2018 at 4:50pm
Kevin: Thank you for responding to my comment. But why was there a “negative credit shock,” as you say?
Tom DeMeo
Jun 6 2018 at 6:44pm
Kevin Erdmann –
I don’t have a wide breadth of exposure to various scenarios and don’t study it like you do, but I do live in one of these places.
My experience has been that municipal governments have been under ever increasing financial pressures. Many of these places have mature infrastructure that is near full utilization.
Towns don’t just want growth. They want high margin growth. Adding a high ratio of residential tax dollars per incremental citizen can often still work. Adding more low margin citizens is highly problematic, particularly when the expensive elements of town governance, such as schools, sewer and water or safety infrastructure may be reaching capacity.
This may not explain a number of other scenarios, but that what I see here.
Kevin Erdmann
Jun 6 2018 at 7:23pm
That’s another good point, Tom. It appears to me that there is a pattern among the problem cities where property taxes on existing property are far too low, and there is an attempt to make up for it by overtaxing or overburdening new development. One way to think about this is that these cities usually have long waits for developers to go through their approval processes, which account for much of the cost, from the developers’ perspective. That is basically a form of queueing, and it is basically pure waste. Developers would happily pay even higher fees and taxes to cut years off of development time. But, these problems didn’t develop rationally. They are a product of the collective action problem of locals having too much power to obstruct real capital investment. So, there is free money just going up in smoke that local polities could claim for those costs, but, really, property taxes on existing properties need to be pulled up to more normal levels. That’s probably not likely to happen. I don’t know the political answer. I’d be happy if we can just stop preventing the school teacher in Topeka from buying a home that is underpriced by 30% today which he can easily afford because we’re afraid of credit bubbles. The central problems in the Closed Access cities may be intractable, but our erroneous reactions to the symptoms of those problems have left a lot of low hanging fruit to pick in the meantime.
Matthew Waters
Jun 6 2018 at 11:05pm
From 2005-08, I can’t fathom an argument that housing was always correctly priced. The 2006 to mid-2008 decline in housing prices and housing starts was in the midst of regular NGDP expectations.
Like with the 1987 crash, housing was either mispriced in 2005 or mispriced beginning of 2008. There is no feasible argument that housing had a rational valuation at both points.
Today’s housing prices are supported by lower interest rates than 2005. 2005 10 year rates were in the low 4%. 10 year rate was below 2% much of the past 2 years. A perpetual cash flow has a PV 33% higher at 3% discount vs 4% discount. A 2% discount is 100% higher than 4% discount.
Matthew Waters
Jun 6 2018 at 11:16pm
A lack of housing oversupply nationally does not make much sense either. Except for a spike of one month in 1984, 2005 housing starts went above every level since 1978. The mid-2000’s had far higher starts than average starts of the 1990’s.
Maybe a demographic argument could be made. But the simple explanation is that credit from subprime loans inducing people to pay too high of prices compared with rents. Price-to-rent nationally is still much less than 2005 peak, despite lower interest rates.
Where prices have recovered, it was due to factors hard to foresee in 2005. The San Francisco area prices boomed in 2005 for similar reasons as the rest of the country. Like the 1987 crash, if the 2005 prices were rational, why were 2007 prices irrational? Prices adjusting on new news, like EMH says, doesn’t validate previous prices without knowledge of future events.
Kevin Erdmann
Jun 7 2018 at 1:41am
Matthew, those are good questions.
Here’s a big reason: Before 2008, mortgage originations generally moved up together across the market. In spite of the apparent spike in subprime lending, there was no spike in lending to low FICO scores, low incomes, low education, etc. In 2008, there was an immediate and permanent collapse in low tier lending. Two-thirds of lending to FICO scores under 720 disappeared and more than half of lending in the 720-760 range disappeared. Above 760, originations continued roughly as they had before. To give you an idea of the scale, 2/3 of households are homeowners and the average US FICO score has been around 690 for the entire time.
I would say that, across the country today home prices are about 20% below previous intrinsic value in high tier neighborhoods and about 40% below value in low tier neighborhoods, compared to the previous longstanding credit regime that predates the bubble.
This isn’t really an EMH issue. We legally removed access to funding for a broad section of the market. That’s a structural shock that changes the price level.
Another way to say it is that, at the margin, large portions of the housing market shifted from owner-occupier to investor-owned, and investors require higher gross yields.
I couldn’t go into all the details in that 7 page policy paper, but the effects of this are quite systematic at the zip code level in every city.
Scott Sumner
Jun 7 2018 at 2:21am
Alan, You said:
“Given that the largest funder of the Center has come out against tariffs, one wonders if there will be a diversity of opinions on that issue.”
It’s pretty hard to find an economist who favors tariffs. To find supporters of tariffs you need to either hire Peter Navarro, or hire a non-economist. (I exaggerate, but only slightly.)
Matthew Waters
Jun 7 2018 at 3:04pm
Kevin,
It’s misleading to look only at FICO scores. The rating agency models of non-conforming MBS looked at average FICO scores. The models were gamed by finding people with high FICO scores but without the incomes or assets also needed to service the loans.
Wikipedia on Gaming Models
I strongly disagree that there was not a spike in poor underwriting for 2003-06 lending. Alt-A and Subprime MBS went to several times previous levels during that time-frame. If underwriting standards did not decrease after 2003, then all of those Alt-A and Subprime loans would have had to have been kept on bank’s own books before 2003.
There is no indication of a similar amount of pre-existing Alt-A and Subprime mortgages, equal to the post-2003 MBS issuance. For example, average LTV of mortgaged houses went from 55% to 90% from 2000 to 2005.
FICO scores since 2008 have been higher than FICO scores before 2000. However, median LTV of new mortgages has also gone from 80% to 95%. FICO offsets higher LTV.
Mortgage Statistics
Bob Murphy
Jun 7 2018 at 4:36pm
Wait a second… Scott, because you put “Bubble” in quotation marks in your title, I thought Kevin was denying that there was a bubble. But actually, he constantly refers to a bubble throughout the paper, including in its title. So unlike you, Kevin is saying yes there was a housing bubble (at least in certain cities), and he’s just offering a different explanation for it, right?
Kevin Erdmann
Jun 7 2018 at 8:14pm
Bob, I share Scott’s discomfort with the use of the word “bubble” here, and I always have the same dilemma of whether to use scare quotes or not.
There were a few cities, like Phoenix and Las Vegas where prices were probably destined to fall to some extent in most counterfactual scenarios, and the use of the word “bubble” is appropriate in that very limited sense. But, 90%+ of what was ascribed to a “bubble” and deemed to lead to inevitable collapse, was clearly not a bubble in hindsight, including housing markets in the half dozen or so cities where home prices were far and away the highest in 2005.
So, in the space of generally accepted consensus of what happened, a small little corner of that space can retain the descriptor “bubble”. The rest of it shouldn’t. There is certainly no reason, for instance, that low end homes in a city like Atlanta should have lost 40% of their value from 2007 to 2012.
The fact that that monstrous outcome is excused as the inevitable outcome of a “bubble” that never was is a much more important fact than the fact that in a handful of cities, home prices could be reasonably called a “bubble” if you need to, for a short time. The difficulty for me is how to talk about the “bubble” (scare quotes) in Atlanta without bringing in all the baggage that people have about the “bubble” (maybe reasonable) in Phoenix.
It’s sort of like the closing scene in Caddyshack, and I’m saying Bill Murray went overboard. I might even say he was fighting an “infestation”. And, you would react by saying, “Oh, so you’re just denying that there was a groundhog. That’s where you’re going to go with this?” It’s a matter of scale.
Or, put another way, any policy prescription that came from identifying 2004-2006 as a bubble was wholly wrong and damaging. Like blowing up a golf course because of one clever groundhog.
Kevin Erdmann
Jun 7 2018 at 9:17pm
Matthew, thanks for the links. My responses are:
This is one of the starting points for my work. If all of these supposedly extreme trends were happening in the mortgage market, then we should see extreme trends in national household data. But the data isn’t extreme. In fact, the data shows nothing. To the extent that there was an increase in debt payments of more than 40% of income during the subprime boom, for instance, it was entirely among households with high incomes, because they were the primary buyers in Closed Access housing markets. Households with lower incomes weren’t buying in those markets in any significant number. In fact, they were moving away from those cities by the hundreds of thousands.
Matthew Waters
Jun 8 2018 at 4:06am
Kevin,
The homeownership data will not give evidence of investment properties, second liens or housing prices becoming too high. Homeownership did increase though, from 64% in mid-90’s to 69%. Subprime loans started in mid-90’s and were unheard of before then.
The poor underwriting was mainly with poor documentation, rather than FICO score or even LTV. Poor documentation and high LTV is ideal for an investor who otherwise has high income and high FICO.
Both the defaults and housing price declines happened *before* Sep. 2008. While NGDP was too tight, the recession at the time was not out of the ordinary. Money was similarly tight in 2001 or 1991. Defaults should not have increased as they did 2007 to mid-2008.
On Open vs Closed Access, Closed Access cities should have both higher prices and higher rents. Price-to-rent increased nationwide, above what was justified by interest rates.
In Open Access cities, the market responded by building a lot of homes. Here in Atlanta, by early-2008, many of those homebuilders went bankrupt. I worked in commercial construction and knew many people who worked residential 2003-07 and then came back to commercial. The market gave a false demand signal due to the investment demand.
One final point: The market mostly acted rational 2004-06. Almost all of the actors rationally maximized their interests. The bag was ultimately held by either bank depositors, Money Market funds or GSE’s. The banks in the middle has much lower risk-weighting. MMF holders were wrong to think that principal would never go down. But other than Reserve Primary, the Fed backed all MMF with unlimited deposit insurance.
Kevin Erdmann
Jun 8 2018 at 10:16am
Matthew,
Someone owns every house. The fact that the private securitization boom was not related to a surge in unqualified owner-occupiers is an important corrective to public understanding. But, it is true that when prices collapsed, leveraged investors were an important cause of rising defaults.
You are right that poor underwriting was mostly in the form of poor documentation. I was as surprised as anyone to find out that there wasn’t any decline in borrower quality associated with it. It is understandable that everyone assumed it had.
Some defaults happened before Sep 2008, but more than 80% happened after. It was the persistence of defaults over years that caused impairments in AAA rated MBSs, to the extent that there were any.
There was no reason for a collapse in Atlanta. Rates of building were about the same from 1996 to 2006. There is no reason to think there was overbuilding. Vacancies never rose in Atlanta, or practically anywhere else. Here is permits in Atlanta, scaled by labor force:
https://fred.stlouisfed.org/graph/?g=k7Dh
Prices in Open Access cities were exactly where they should have been, given the low interest rates. Price/Rent in Atlanta was well within the long term range. Your friends were put through an unnecessary correction. They should still be in residential construction.
Matthew Waters
Jun 8 2018 at 5:42pm
Hmm, that’s interesting about the trends in Atlanta homebuilding. The major part of the declines did happen before Sep. 2008.
Even more interesting, Atlanta had a later price peak than the nation generally. Prices peaked in July 2007. 50% of the decline in housing starts happened before the price peak. I have some theories, but they would be speculative.
On MBS writedowns, 10’s of billions of writedowns happened before Sep. 2008. HSBC had a $10 billion writedown in Feb 2007. The fact that AAA tranches suffered partly due to the Great Recession does not exonerate the AAA rating. AAA assets were considered strong enough to survive a Great Depression scenario.
Kevin Erdmann
Jun 8 2018 at 5:56pm
Matthew, here’s another piece of the puzzle for you. Between the end of 2005 when housing starts started falling off a cliff, and the end of 2007, rent inflation in Atlanta went from zero, where it had been throughout the boom, to 7%! Kind of strange for a city supposedly suffering from oversupply.
https://fred.stlouisfed.org/graph/?g=k7RD
Then, from the end of 2008 to the end of 2011 we solved the rent inflation problem, temporarily, by destroying working class balance sheets and evicting thousands of Atlanta homeowners who then had to massively cut back on their housing consumption.
There’s only so much you can do there, so now, rent inflation in Atlanta is just chugging along at 5% because we have basically outlawed entry level homeownership.
You don’t need to speculate. I’m telling you the story. 🙂
On the MBSs, the write downs were based on market values. Basic AAA MBSs generally performed quite well, especially if you consider the interest rate gains they had. Only about 10% suffered any actual impairments to their cash flows. CDOs are another story.
Matthew Waters
Jun 9 2018 at 7:47pm
That’s what I find strange. The market for Atlanta was still healthy, but residential construction fell off a cliff before Lehman. Loans to homebuilders in 2007 should have remained strong.
My speculation was that housing construction loans are looked at nationally by lenders. So CA, NV, AZ and FL construction defaults hurt GA loans.
Also, bubbles can have an overcorrection effect. The NASDAQ became underpriced in 2002. Some South Korean stocks became very underpriced after the 90’s financial crisis. Buffett talked about buying net-nets in SK, which hadn’t been around since the 50’s.
Finally, the new homes built in Atlanta could have been mismatched to the demand. While Atlanta-area prices and rents stayed strong, *new* homes in particular in 2007 may have struggled. Housing was not necessarily fungible.
Kevin Erdmann
Jun 10 2018 at 11:44am
If you just remove the idea that there was a bubble in Atlanta, all the mysteries and the need to speculate about what happened go away. The idea that there was a bubble is what is making this complicated to you.
The market in Atlanta was healthy, then we imposed a series of monetary and credit shocks on it. It’s what you know that just ain’t so that’s tripping you up.
Willy2
Jun 10 2018 at 12:46pm
– Scott Sumner doesn’t believe that bubbles exist. He can’t identify a bubble even when it has written “bubble” all over it. Instead he believes that market are (ruthlessly) efficient.
– The problem is that both notions are true at the same time. Markets are (ruthlessly) efficient and can be in a bubble at the same time.
– A good example currently is Sydney. Homeprices over there are at 12 times gross (average ??/median ??) household income. A ratio of 4 to 4.5 times gross income is already considered to be “too high”. 12 times income is clearly “a bubble” by ANY standard. (But we won’t hear one S. Sumner about that right ?)
– But why is real estate in Sydney at 12 times gross income ? Because “efficient markets” have pushed prices this high. Australian banks were (past tense) willing to lend people 12 times their gross (household) income. Australian banks are now reducing the amount of credit they’re willing to lend to (far) below 12 times gross income. And that will deflate the australian housing bubble.
Willy2
Jun 10 2018 at 1:43pm
– @Kevin: Atlanta was in 2007 not in a bubble ?
– Real estate in Atlanta peaked in 2007, it fell some 35% between 2007 and 2012. The same real estate index is now some 5% above the peak of 2007. Not in a bubble ?
– “Dodd-Frank” was passed in 2011, 4 years AFTER real estate in Atlanta peaked (in 2007). And one year before real estate started to recover. Then one could argue that Dodd-Frank actually helped to recover the real estate market(s).
– The article below shows how real estate prices have recovered after say 2011. and some market are actually ABOVE their 2007 peak. Some segments of the markets may not have recovered but overall the Case-Shiller index is above their 2007 highs.
– To explain why there was a bubble one has to look at CREDIT. (See my previous post). When one throws A LOT OF credit (e.g mortgages) then yo will see real estate prices rise. And that creates the ILLUSION that there’s an undersupply (of e.g. real estate. (See my previous example of Sydney).
Source:
https://wolfstreet.com/2018/05/29/update-on-the-most-splendid-housing-bubbles-in-america
Kevin Erdmann
Jun 10 2018 at 4:16pm
Willy, both prices and credit outstanding will rise in either a positive demand context or a negative supply context. The appearance of either tells you nothing about causality. In the comments above, I laid out some better indicators of causality. The passionate reliance on poor indicators is what got everyone so off track. And Atlanta may be the single best example of a city that was hammered by Dodd-Frank.
Comments are closed.