Back in 2009-10, I did a number of posts criticizing the theory that rising house prices in the early 2000s represented a “bubble”. In one post, I pointed to an article in The Economist that criticized Eugene Fama, and bragged that they had presciently foreseen the housing bubble. In fact, the specific predictions they cited (from an 2003 advertisement for The Economist, since deleted) turned out to be almost entirely wrong, indeed wildly off base.
The Economist did not take kindly to my post:
Mr Sumner disagrees. He seems to think it’s funny that The Economists pent much of the last decade warning that, globally, home prices were rising in a troubling manner. Contrarianism is fun and all, but this strikes me as an odd way to process the experiences that led us to this point.
I would note that Free Exchange seemed to enjoy making fun of Fama’s views.
Now The Economist has seen the light:
Perhaps it is just a matter of time before the house of cards collapses. But as a recent paper by Gabriel Chodorow-Reich of Harvard University and colleagues explains, what might appear to be a housing bubble may in fact be the product of fundamental economic shifts. The paper shows that the monumental house-price increases in America in the early to mid-2000s were largely a consequence of factors such as urban revitalisation, growing preferences for city living and rising wage premia for educated workers in cities. By 2019 American real house prices had pretty much regained their pre-financial-crisis peak, further evidence that the mania of the mid-2000s was perhaps not quite so mad after all.
Fundamental forces may once again explain why house prices today are so high—and why they may endure. Three of them stand out: robust household balance-sheets; people’s greater willingness to spend more on their living arrangements; and the severity of supply constraints.
I’d add permanently low real interest rates.
I would like to take credit here, but I was just shooting from the hip when I questioned the housing bubble view that was so popular after 2006. Credit should go to Kevin Erdmann, who produced a mountain of evidence against the bubble hypothesis in two very impressive books on housing. His view, which was once highly contrarian, has now been completely vindicated. Indeed, I don’t see how any fair-minded person reading his books could still believe in the housing bubble theory. Unfortunately, he’ll probably be ignored. The media tends to focus on academic research from top schools like Harvard, not unaccredited individuals working on their own. Better to be famous than to be right.
READER COMMENTS
Todd Ramsey
Jan 23 2022 at 12:40pm
I’m not trolling, just seeking understanding:
What is your framework for why you think interest rates are permanently low?
What is your estimate of a “normal” (during a hypothetical time when the Fed is neither buying nor selling Treasuries) 10-year treasury rate going forward?
Scott Sumner
Jan 23 2022 at 5:37pm
Real interest rates have been trending lower for four decades. I’d expect the 10-year Treasury yield (in real terms) to be zero percent or lower going forward, even under a neutral Fed policy.
Brent Buckner
Jan 24 2022 at 8:23am
One could even refer to the 800 year trend!
(c.f. https://www.bankofengland.co.uk/working-paper/2020/eight-centuries-of-global-real-interest-rates-r-g-and-the-suprasecular-decline-1311-2018 )
Mark Brady
Jan 23 2022 at 2:49pm
A quick thought. Why can’t both ideas–that from time to time there are housing bubbles, and that there is a long-run trend toward higher house prices–be compatible for at least some periods in some countries. I’m thinking particularly of the U.S. and the UK.
Scott Sumner
Jan 23 2022 at 5:39pm
That’s theoretically possible, but I see absolutely zero evidence in favor of that hypothesis. Zero.
Bubbles are simply not a useful hypothesis. The theory doesn’t provide useful predictions.
Cranmer, Charles
Jan 23 2022 at 7:02pm
Hello Mr. Sumner! I just wrote a comment on your excellent piece on MMT when I saw this article. It’s been a long time since the Financial Crisis and I am still screaming in the wilderness to get the world to understand what really caused it. It was not caused by free markets run amok. IT WAS CAUSED BY IDIOTIC BANK REGULATIONS: THE BASEL CAPITAL STANDARDS. SEE MY BLOGPOST “BASEL: FAULTY.”
https://cantercap.wordpress.com/2016/06/07/first-blog-post/
These regulations drove European banks and US shadow banks to become egregiously over leveraged and illiquid by booking gargantuan quantities of AAA sub prime MBS, financed with short term repo and Asset Backed Commercial Paper. Basel considered the MBS “low risk:” requiring minimal capital.
If everyone accepted that the crisis was caused by bad regulation, it might humble some of the more virulent advocates of regulation, and might get Elizabeth Warren to finally shut up.
Please take a look at my piece and let me know what you think.
I am not familiar with Kevin Erdman’s work, but I’ll take a look
Jonathan S
Jan 23 2022 at 8:05pm
What are your thoughts on using the home price:household income ratio to evaluate whether home prices are “too high” or “too low”? (Mortage payment:income ratio may be better, but I cannot find good data for that.) For example: https://www.longtermtrends.net/home-price-median-annual-income-ratio/
I don’t love the term “bubble” either since it seems conjures an image of something formerly of value ceasing to have value. As long as humans aren’t extinct real estate should still hold value.
Kevin Erdmann
Jan 23 2022 at 8:31pm
Jonathan,
The problem is that it is very difficult to come up with any generalized US housing measure that is informative because the US housing market has become so overwhelmingly driven by local supply constraints that the variance between cities is huge. Even within metros, the variance is huge. In LA, for instance, high end neighborhoods might have price/income ratios of 6 or 7, while at the low end it is well over 10. In a city like Dallas, it is relatively similar through the metro. I’d guess there now it’s around 3 or 4. Certainly, taking liquidity out of housing markets can reduce the p/I ratio in low end LA. There is some correlation. But mostly the ultra high price/income ratios are signaling extreme supply constraints, and by the time they get thrown in the blender to come up with a national average, they are more likely to mislead about potential demand factors than to add information, imho.
David S
Jan 24 2022 at 7:53am
I just ordered Building From the Ground Up–buy Scott a cup of coffee someday–or a sandwich from Subway.
Thank you for taking a deeper dive into the rank swamp of American housing, particularly by pointing out the granularity of housing markets. In the Boston suburbs I know real estate agents who specialize in neighborhood markets to such a degree that they refuse commissions in surrounding towns. That’s a high-end phenomenon, but the principle holds in many urban areas.
The only reliable generalization is the persistence of supply shortages in the ultra-regulated coastal regions. I see no solution to this in my lifetime–and for better or worse I make a living off the scarcity associated with this problem.
BTW, did you see the Ronan Lyons post that Tyler linked to at Marginal Revolution?
Kevin erdmann
Jan 24 2022 at 11:59am
Thanks David. I hadn’t seen that post on Ronan Lyons.
nobody.really
Jan 24 2022 at 4:41am
HEY–Erdmann should respond as Dr. Science would: “I have a MASTER’S DEGREE…!”
Kevin erdmann
Jan 24 2022 at 12:00pm
😅
Scott Sumner
Jan 24 2022 at 12:20pm
Those snobs won’t accept anything short of a PhD. 🙁
SK
Jan 24 2022 at 8:40am
Just wondering about the degree to watch housing prices would have been lower had banks been holders of mortgages and no securitization markets had existed. Also, as best I remember a low yield environment had all around the globe searching for what they hoped were better risk adjusted returns, or just higher returns via securitized bond buys at whatever tranche they bought of the bond issuance.
Banks clearly pushed product out the door and gave little thought to credit quality, instead relying upon rating agencies for the ok and they in turn seemed to rely upon questionable underwriting criteria. Seems the whole financial system was geared to those involved just interested in fee revenue.
Scott Sumner
Jan 24 2022 at 12:22pm
Kevin has a lot to say on this. I’d encourage people to read his books.
Phil H
Jan 24 2022 at 9:11am
Sorry if this is a tired debate, but a couple of things struck me.
First, the US economy is 50% bigger now than it was in the mid-2000s. Are these prices adjusted for that?
Second, one quite reasonable way of understanding the word “bubble” would be: prices are likely to fall a lot in the near future. And they did indeed fall. It would seem reasonable to say something like: a lot of people were using set of assumptions about what housing “should” cost that were no longer realistic, and so their claims of a “bubble” were inaccurate. But if you bought an expensive house in 2005, you lost your shirt for a decade. That sounds exactly like a bubble.
Scott Sumner
Jan 24 2022 at 12:23pm
But in most countries prices did not fall from “bubble” levels, at least to any significant extent.
Scott Sumner
Jan 24 2022 at 12:30pm
Check out this graph:
https://twitter.com/binarybits/status/1485635627890716673
Jon Murphy
Jan 24 2022 at 12:29pm
That sounds like reasoning from a price change to me. Prices can fall for any number of reasons. For example, at the beginning of the pandemic, I predicted that oil prices would fall significantly due to the lockdowns. That indeed happened, but the fact oil prices fell doesn’t imply oil prices were in a bubble before.
MikeDC
Jan 24 2022 at 1:25pm
It seems to me that if a Bubble is to mean anything, it has to mean an increase in demand driven by speculative purchases.
Are people buying houses to live in, or are they buying houses because they believe they can flip them and resell them for more a year down the road.
Anecdotally, the folks who bought my house in NoVA in 2005 thought the latter. Various studies, such as this one: https://www.princeton.edu/~wxiong/papers/Speculation.pdf
have indicated that my anecdote was part of a larger trend and that there was a substantial increase in the level of speculative purchases during this time.
It sounds incorrect to say that there was both a housing bubble and a housing shortage, but I don’t see why they should be mutually exclusive when we know regulatory policy greatly restricts supply and subsidizes demand.
Jon Murphy
Jan 24 2022 at 2:42pm
Maybe, but that would involve any purchase with a temporal element. For example, ahead of major storms, one tends to see prices rise as people aim to stock up on essentials. That’s an increase in demand driven by speculative purchases. Does that imply that pre-storm prices are a bubble?
jim
Jan 24 2022 at 3:11pm
Price surges ahead of storms result from weather forecasts, which are overall pretty accurate, so it’s probably not appropriate to call storm stocking “speculative”. Also people aren’t stocking to resell at a higher price, they’re stocking against the likelihood of need.
MikeDC
Jan 24 2022 at 3:52pm
I thought about that… maybe, but I think pre-storm prices can be distinguished because I mean “speculative” in the technical sense of buying a product not because you want to use it, but because you want to resell it for a profit.
In the case of a storm, most of what’s happening is people consuming based on the belief that a supply restriction is coming. For example, I buy toilet paper because I think I won’t be able to next week when we’re shut in by the storm.
Just for bonus points, I would also point out that presumed “price gouging” isn’t strictly speaking, speculation either. So if I hear about a storm coming to Florida, buy toilet paper, and drive down to Florida to sell it for a profit, I’m actually “producing” more toilet paper and introducing it to the market.
The true speculative case would be that I go in, buy all the toilet paper in the store, and then re-sell it to my neighbors. That would, I guess, be a bubble and it does happen from time to time, but most of the time it’s a really minor effect compared to the other two.
nobody.really
Jan 24 2022 at 4:52pm
It does? Good to know. I’ll be dropping by the next time I can’t find toilet paper.
Jon Murphy
Jan 24 2022 at 6:00pm
That would then imply all commerce other than manufacturing is speculation as all firms buy with the intent to resell at a higher price. Consequently, all prices would be bubble prices.
MikeDC
Jan 25 2022 at 9:13am
That would then imply all commerce other than manufacturing is speculation as all firms buy with the intent to resell at a higher price. Consequently, all prices would be bubble prices.
Nope, not at all. Consider the example I gave of folks to who buy toilet paper in a low price place and transport it to a high cost place. That’s not speculation, that’s supply.
This is a pretty fundamental ECON 101 definition. Supply is almost always speculative but not all speculation is supply. I get that there’s some nuance in individual cases, but generally speaking, when we’re talking about “speculation”, we are talking about speculation absent any obvious act of supply/production.
Mark Z
Jan 24 2022 at 3:09pm
While this is true, I think the housing shortage played a much bigger role than easy lending. Ed Glaeser and co-authors wrote a paper on this (be forewarned, there’s some math) finding that only a small fraction of the alleged ‘boom’ could be explained by cheap credit, I think ~20%. The insinuation I took from it was that most of the price increases were supply and demand.
vince
Jan 24 2022 at 3:45pm
Great link to the 2010 Glaeser et al paper. Yes, it attributes 20% of the price increase from 1996-2006 to interest rates. It also found no evidence of a role played by loan approval rates and loan-to-value ratios, while admitting that those factors need more research and better data. It also doesn’t pursue an argument by Shiller that mass psychology is more important than those credit issues.
Aladdin
Jan 24 2022 at 1:43pm
That implies a sudden collapse in demand for housing in the 2008 crisis. What exactly caused such a high and sudden drop in demand / rise in default rates?
Thats the question I have anyway, what exactly caused this, if not a bubble?
Kenneth Duda
Jan 24 2022 at 2:07pm
Aladdin, the shortest answer is “please read Kevin’s books”.
The short answer is “the Fed kneecapped the economy in 2008, people lost their jobs, couldn’t pay their mortgages, and everyone had to sell at the same time, causing prices to go down in 2008-9 only to come back up more strongly after, which is the opposite of a bubble — it was a temporary depression.”
The less-short answer is:
1. House prices rose through 2000-2008 due to supply constraints
2. In 2008, there was a financial crisis due to a run on a shadow banking system that was based on short-term borrowing collateralized by real-estate-backed securities — when people became worried about the quality of the mortgages underlying those securities, liquidity froze up, and investment banks arbitraging term premia couldn’t roll over their debts, and faced illiquidty (short-term insolvency)
3. The Fed decided to bail them out through TARP — creating money and simply buying high-quality-but-illiquid mortgage-backed securities (a program that profited them billions of dollars BTW)
4. The Fed became concerned that the money created in TARP would overheat the economy, and tightened monetary policy, kneecapping the economy
5. The tight money caused the economy to collapse, leading to excess unemployment (thanks to sticky wages), which led to foreclosures (because people lost their jobs)
6. All the pundits lost their s*** about how obviously home prices had been way too high and the whole thing was.a “bubble” “fueled by” “cheap money”, blah blah blah
7. After the economy recovered from the Fed’s kneecapping, house prices zoomed past their 2008 highs, but somehow, people still think there was a bubble. When you look at the actual history of house prices, they rise steadily from 1970 to 2022, with a ditch in 2008-9. This is pretty much the *exact opposite* of a “bubble”.
Kenneth Duda
Menlo Park, CA
vince
Jan 24 2022 at 2:31pm
Great less-short answer. On item 1, the shortage was probably worsened by low interest rates, lax lending standards, and the government push for homeownership. Like MikeDC said, a bubble and a housing shortage are not mutually exclusive. Items 3 and 4 describe how unfair the bailout was. Bankers like Goldman Sachs took risks and lost nothing on those risks when AIG was bailed out, allowing them to completely pay Goldman and others. As many have called it: private profits, social losses. On item 6 and MikeDCs bubble comment, the simplest definition I’ve heard is you get a bubble when price increases beget further price increases.
Kenneth Duda
Jan 24 2022 at 3:16pm
Hi Vince, I agree with most of what you said. In particular, I agree the bailout was extremely unfair.
I’m not so sure about your first point:
Low interest rates would actually help alleviate the shortage (by making it easier to finance new construction) if only dense construction were legal in desirable areas. Lax lending standards did not actually exist, but if they had, they could also have helped with the shortage for the same reason. Finally, I’m not quite sure what you are referring to specifically in the phrase “government push for homeownership”, but if you’re referring to the deductibility of mortgage interest, I can’t see how subsidizing housing construction could lead to a housing shortage.
Maybe you meant, “given that housing construction is mostly illegal, high housing prices were probably worsened by … ” ?
In my view, the housing shortage has a very simple cause: we prohibit building more housing in desirable areas, and have for many years. Because of this supply blockade, a slowly increasing population collectively has to reduce per-capita real housing consumption (pigeonhole principle), which happens only when the price of the housing people want rises out of reach (rationing via market pricing). It’s such a sad and crazy situation. We are collectively choosing deprivation over abundance, and making up nonsense stories about corrupt banks and evil developers and housing bubbles to avoid facing the reality that we are just doing this to ourselves for no good reason.
-Ken
vince
Jan 24 2022 at 3:59pm
Ken: On low interest rates, I refer to an increase in the demand side. The Glaeser paper for example attributed 20% of the price increase to it. Some examples of lax lending standards are no-doc (liar) loans, subprime lending, and teaser ARM rates. The homeownership push started with the Community Reinvestment Act.
David Seltzer
Jan 26 2022 at 6:24pm
Kenneth, thanks for the concise explanation. I wonder how restrictive zone regulations affected supply in terms of new housing construction. It would be interesting to compare rates of price increases against strict/relaxed zoning regulations in randomly selected urban housing markets. Again, thank you.
Lawrence D Glenn
Jan 24 2022 at 10:34am
I can’t help but wonder what part of the housing equation is related to currency debasement.
Kevin Erdmann
Jan 26 2022 at 9:41pm
This currency?
https://fred.stlouisfed.org/graph/?g=LgMM
Ted Durant
Jan 24 2022 at 3:33pm
The “Newtonian Physics” of Housing are pretty simple: slow-moving supply, fast-moving demand. The bubble aspect of anything is that excess price growth generates its own speculative trading demand. When change in supply is slow to react to changes in demand, you have conditions that are ripe for speculative trading activity. For example, a bubble in toilet paper and hand sanitizer in 2020. So, you think bubble is an incorrect term? Fine, how about persistent disequilibrium in a market? Disequilibrium in housing markets tends to last a long time and, in fact, might be described as the normal condition of most housing markets. Happily, disequilibrium in the TP market didn’t last as long.
It is simply incorrect and misleading to say the home prices didn’t fall from their peak in 2006. They fell a lot – nationally more than they did in the Great Depression, maybe more, depending on your choice of index. Sure, they have since rebounded and grown back to peak levels. That does not invalidate the idea that there was a bubble (persistent disequilibrium) in 2004-2006. As long as asset values can be expected to increase over time, they can be expected to exceed some peak bubble value eventually. Whether someone has overpaid for the asset at any point in time depends ultimately on what happens to the price subsequent to the purchase. If you buy an asset expecting returns from increased prices, and 20 years later its price is the same as when you bought it, you overpaid. Of course, housing is a combination of investment and consumption, so the definition of “overpaid” has to be taken with some suspicion. We assume generally that nobody thinks they are significantly overpaying when they sign the purchase offer in an arms-length transaction, and if the buyer lived in the home they got 20 years’ worth of housing consumption out of it.
Another aspect of housing that seems to be either ignored or misunderstood is that money and housing are complementary goods. Around half the people who own single family homes, and the large majority who purchase them in any period, had/have to borrow (i.e. buy money) to do so. And, everyone who has equity in a home and isn’t credit impaired has the ability to borrow (buy money) at a lower rate than almost everyone who doesn’t have a home. (Current rules allow for around 20:1 max leverage of single family residences.) Monthly mortgage payments are a terrible way to measure the cost of housing because they are a combination of the cost of housing and the cost of money.
Home prices themselves are a mix of land (which is generally an appreciating asset) and structure (which is generally a long-term depreciating consumption good). A few people have tried, but nobody has yet successfully published an established index that disaggregates those prices. The indexes that are generally used and bandied about in this discussion are weighted repeat sales indexes, built on price changes observed on the same property over time. A change in preferences for type or location of housing should have no impact on prices measured this way, as increases in price of the more desired factors should be offset by decreases in price of the less desired factors. The reality is more complex, however, as it depends on whether the factors relate to land (“they aren’t making any more”) or structure. Importantly, it’s also likely that properties in demand sell quickly and properties out of demand linger unsold for a long time, which would cause a WRS index to show an overall increase in prices in that period.
Summary:
2001-2007 was a period of significant increases in demand (meaning a shift of the demand curve) for single family housing, driven by two components: 1) credit-constrained households for whom money had not been available at a reasonable price were now able to borrow; and 2) speculative trading.
2008-2013 was a period of significant decreases in demand as the credit bar was once again raised and the speculators died/fled the market (though different kinds of speculators entered the markets, as well).
2013-today demand has increased significantly again, as a very large population cohort of first-time home buyers began entering the market and the credit bar has been lowered (though, nowhere near as low as 2007). Notably, there has been a significant shortage of properties available for purchase in that segment.
2020-today – Rapidly changing housing preferences as a result of COVID-19 and related changes in work location requirements.
Michael Sandifer
Jan 24 2022 at 4:18pm
I agree with all of this, except for this sentence:
“I’d add permanently low real interest rates.”
Low interest rates, permanent or not, do not inflate asset prices. To state otherwise, is to contradict the Sumnerian principle that interest rates are an effect, not a cause.
For an example, just look at Japan.
I understand that in finance models, real interest rates are a variable in asset valuation. I think that’s wrong. What really matters is current versus expected NGDP growth. When rates are low due to low current NGDP growth versus higher expected growth later, asset prices will be relatively high.
Hence, in the 90s and early 2000s, stocks were not overvalued, and there was no stock bubble. Monetary policy was simply tight and tightening. As expectations for ever tigher money moved out into the future, the economy slowed until current growth turned negative, despite a historic productivity boom, and one that was originally expected to get even bigger going forward. Then, a combination of many companies killed in their infancy by tight money, and the continuation of tight monetary policy, led to a very long period of recovery, particulary for NASDAQ stocks. I think the killing of the tech boom probably caused significant supply-side damage to the economy and helped lead to more harmful consolidation in tech than would have otherwise occurred.
The same is true of the real estate market over this period, and on through 2005, except that there were also supply-side factors at work, such as restrictive building laws.
robc
Jan 25 2022 at 9:32am
I have disagreed with Scott in the past about the housing bubble. I think it was a bubble. I will attempt to define bubble in a way that makes sense to me. Feel free to pick apart this definition.
Bubble – an extreme temporal arbitrage opportunity.
I think bubbles are rare, but they do exist. Obviously arbitrage across time exists. Where are the bubbles? The largest ones.
John Paulson arbitraged to a gain of $20B. Sounds extreme to me.
Kevin Erdmann
Jan 26 2022 at 9:56pm
John Paulson’s gain was speculative.
A better example of arbitrage was buying up AAA mortgage securities in 2008 that were in technical default but that never ended up having impaired cash flows.
Comments are closed.