I.
have long believed in a paradox of popular narratives. When multiple contradictory narratives emerge around a topic of import to broader society, they all contain parts of the truth (fringe conspiracy theories notwithstanding). But when a dominant society-wide narrative emerges it is usually fundamentally false. For instance, when crime was precipitously falling from the mid-90’s to the mid-20-teens, poll after poll showed that up to 90% of (bipartisan) Americans perceived crime rates at least level over that period, with a majority seeing them as getting worse.
There was a whole genre of popular explanations for the housing “bubble” and Great Recession. Though they varied on the specifics, they all shared the same familiar premise: there was a speculative bubble in real estate fueled by “irresponsible” lending which resulted in over-building; when it popped, as all bubbles inevitably do, leveraged banks became insolvent and either failed or needed propping up; the aftermath led to one of the deepest recessions in living memory.
In these polarized times when we as a society cannot agree on very basic facts, one would be hard pressed to find anyone who lived through that time and would not be on board with the above description of the economic crisis (what should have been done about it is another story). But in his book Shut Out, author Kevin Erdmann persuasively argues the dominant narrative that has emerged is incorrect, and rests on a series of myths that appear superficially true, but collapse under careful scrutiny.
His thesis is that the run-up in housing in the mid-2000’s was not an asset bubble superimposed on the crest of a cycle, but a regime shift to higher housing prices caused by supply restrictions in the most economically prosperous and dynamic US cities, which he has deemed “closed access cities” (and include NYC, LA/SD, Silicon Valley, and Boston). “Loose” credit was a necessary facilitator of people moving into and out of these cities and their unique housing markets, was more effect than cause of increased housing prices, and did not focus on buyers with abnormally low credit rating or income. People left these dynamic areas preferentially for more affordable housing in rapidly growing areas in the Southwest and Florida, deemed “contagion cities” (including Las Vegas, Phoenix, inland California, and Tampa). And finally, the subsequent bust and recession were not foreordained consequences, but were imposed by monetary and regulatory policy, with various levels of intention. These first triggered fear in the housing market ending the boom, then allowed a deep recession to proceed that set off a financial panic, and finally, resisted a recovery in the housing bust which had by then spread to the rest of the country and would last well into the next decade.
In short, there wasn’t a demand bubble, but a supply problem; the banking shenanigans were a bit player; and the federal reserve misidentified the problem and responded with a series of disastrous decisions.
II.
These are pretty audacious claims. Ben Barnanke is seen by many as a hero who competently steered us way from an even deeper crisis. The Big Short was a best-selling book made into an Oscar-winning film. Major legislation and new federal entities were created wholly on premises that were the antithesis of the conclusions of Shut Out. Occupy Wall Street, to my knowledge, never tried to Occupy the Zoning Board.
So, Erdmann has to really deliver the goods, and deliver he does. It is one of the more relentless data-driven exercises in upending conventional wisdom I’ve encountered. It simply notes that the credit-fueled bubble thesis necessarily implies certain facts, and then looks at local markets individually to see if the data support that story. And time and time again, they do not.
There are basic facts that would have to be true to support the conventional story: 1) the large price increases should be widespread throughout the US; 2) given that mortgage-credit markets are not international, the US run-up in prices would have to be unique; 3) homeownership and home-building should be unusually high; 4) mortgage securitization should have led or coincided with homeownership; 5) the quality of borrowers and mortgages (credit and income) should have decreased; 6) the initial bust should have hit the low-income owners the hardest; 7) foreclosures should have happened soon after the bust; 8) and given the regulations that arose from post-crisis legislation, the prices should not have risen again to those heights. None of these are accurate.
(1) The price increases that made headlines were a hyper-local phenomenon. Erdmann only looked at twenty metropolitan areas, but among them, only those in the “closed access cities” on the California and northeastern coasts and the “contagion cities” in the Southwest and Florida did prices even resemble a bubble. In the other metro areas he surveyed, and in the suburban and rural areas of the country he did not, price increases were modest or non-existent.
(2) The same price behavior was seen in all English-speaking countries in the early/mid 2000s, and all except the US and Ireland saw a quick recovery to that high price trend after the recession (Australia's was unabated because they even avoided the recession). To be sure, the US was unique, but it was in the price behavior after the bust, not in the boom, that stands out. US policy- and credit-centric stories fail to explain the housing markets in Australia and the UK because there wasn’t really a “bubble”.
(3) Homeownership, adjusted for age, was not increasing; new housing, adjusted for population, was being produced just above its long term average, nowhere near its high. These are often confused, because of composition effects. Homeownership rates did increase several percentage points, but this was a function of an aging population. Single unit-housing starts were at a peak, but this was partly due to a permanent shift away from multi-unit apartments and manufactured homes. Total housing units were being produced at a very modest rate given the increased prices.
(4) Mortgage securitization held by private institutions (which represent subprime loans) severely lagged the increase in homeownership, and first-time homebuyers as a percentage of all homebuyers actually decreased during the period from 2003-06 where these loan products took off.
(5) The marginal buyers coming into homeownership had high income (see more in section below), not low. At each income quintile, leverage was decreasing.
(6) During the initial bust, the rise in delinquencies was concentrated in the population with high income and credit scores; many were investors who were strategically defaulting.
(7) Vast majority of foreclosures happened well after the bust because they were due to the larger recession (and occurred in areas that did not see a run up).
(8) Prices have resumed their upward trend in the face of a prolonged credit bust in areas that had seen the biggest price increases during the boom, because it was fueled by supply restriction.
III.
It would be impractical to relay all the data that support Erdmann’s thesis, but I’ll go into one example to illustrate. Again, the conventional story is that credit was expanded to underqualified homebuyers who could only justify a given mortgage because home prices were going up. Clearly, if this were true you would expect to see marginal buyers in places where home prices were increasing the most be lower-income. But this is not the case.
Net domestic migration in the “closed access” metro areas between 1995 and 2008 was a) negative, and b) the mirror image of that seen in the “contagion cities.” At the height of the housing boom in 2004, these closed access cities lost 1.5% of their populations, and the contagion cities gained 1.5%; afterward, these trends reversed as housing prices rose more sharply in the contagion cities, and then prices everywhere fell during the recession.
This relationship between the migration statistics in the two groups is not a coincidence. (Shocking facts: if you were a resident of Boston or New York around 2000, and you moved to a newly built home by 2006, you were as likely to be living in Florida as to to still be living in NY or Boston. More homes were built in AZ/NV/OR just for former residents of SF/LA than were built in SF/LA at all). And the timing of the changes in each category are a major hint that there was no speculative bubble. Erdmann writes, “ Doesn’t it seem strange that every location supposedly caught up in a speculative frenzy is characterized by thousands of residents moving away as home prices rose? If the problem in the Contagion sites was too many homes, why would these cities see a sudden drop in in-migration and an increase in out-migration [in 2006]?”
And this sudden shock to inter-city migration in 2005-6 explains the pictures you probably saw of empty homes and neighborhoods. For years, Phoenix had been issuing building permits at an unprecedented pace, but this barely kept up with demand from rising in-migration, as prices still rose significantly. The migration shock was sudden, akin to the music stopping in a game of musical chairs. But in this game, chairs were appropriately being manufactured because the number playing in the game had been increasing, and when that stopped (temporarily), there was a temporary overhang in these markets.
Census data demonstrates that the population of people that had been leaving the closed access cities (and entering the contagion cities) consisted primarily of two groups: homeowners with well-above average incomes who were cashing in their increased equity, and renters with below average incomes who were escaping to affordability (and were still renting, meaning these new units were investor-owned properties). The people moving into and buying in the closed access cities also had above average incomes (and high educational attainment).
None of these groups that were on the move in this period, either of the high income populations or the investor-landlords, fit the stereotypical picture of an unqualified buyer who might be subject to a “predatory” loan during a credit boom. Indeed, the aggregate number of homebuyers from the observed cities in the 2nd and 3rd income quintiles actually decreased from 2003 to 2005 (the boom peak), while high-income homebuyers increased.
This is the story of a broken housing supply reshuffling the deck, not of irresponsible lending to marginal households.
IV.
Even if most people could be convinced that there wasn’t a classic bubble in housing, it’s an even taller order to convince them that the larger financial and economic crisis wasn’t a direct effect of the housing boom-and-bust.
Erdmann devotes much less time to the connection between the crises, but I think it would have been better to dive in those deeper waters (though, in fairness, in doing so, he would have been leaving behind his original contribution and getting into topics that are contentious, and probably irresolvable, among professional economists). He later did so in a paper with Scott Sumner laying out the thesis, and I find it just as convincing.
But, even in Shut Out, he provides enough evidence to be convincing in its own right. It shouldn't take much, because if there was no bubble, there didn’t have to be a crash; and it’s entirely plausible that whatever caused the housing crash could spread to the larger economy. Time and time again, the federal reserve, politicians, academic economists, and the financial media hammered home the message that there was an irrationally overheated housing market, the correction was coming, and it was a matter of just when and how bad.
Markets are forward-looking and all relevant data matter, and the housing market is no different. From 2004 until 2006, two things happened - the federal reserve had begun raising rates and Bernanke et al. had begun cheering on declining home prices and construction. It’s no surprise, then, that is exactly what happened. The market could see the writing on the wall. There was an ensuing increase in mortgage defaults due to rising prices and higher interest rate resets.
But this was likely manageable and would not have been a big deal by itself for mortgage backed securities and CDOs; the problem that led to a panic is that policy makers showed no signs that this was going to stop because they had assumed it was all necessary. Once that panic set in, and because the Fed was so laser-focused on housing, they missed many economic signals of a pending recession (most notably, collapsing projected nominal GDP), with a disastrous series of decisions to follow.
So, no, the housing bust did not cause a financial crisis that led to a recession. It is more accurate to say that the same forces - bad monetary policies - caused all three. And then for the cherry on top: post-recession legislation severely tightened credit markets which persist to this day. This turned a deep recession into a deep recession plus a long, slow recovery, as the housing and credit markets in most of the US still haven’t returned to 2000-levels. Homeownership rates for Generation X and younger in 2017 were the lowest they’ve been since before 1982. Housing prices in low-income areas outside of major cities, which did not participate in the price boom, nonetheless saw the largest decreases in the decade to follow.
Erdmann: “There is an awkwardness...in the credit bubble theory...which was solved by invoking behavioral finance. Buyers and lenders were irrational, and a collapse was unavoidable… A model whose very specifications are that buyers and sellers are unmoored from objective reality is difficult to falsify. So, as home prices rose higher and higher in 2004 and 2005, even as the Fed raised...rates, the model was adjusted to account for buyers and leaders who must have been even more irrational than previously thought...The problem is that, if a behavioral model influences policy, then collapse is imminent whether the model is right or wrong - if not because of agent irrationality, then because of model error. Those who ‘called’ the bubble have regarded the collapse as confirmation that there was a string or irrational players… What if model error in public policy was the primary cause of the collapse?” Indeed. (Emphasis mine.)
(None of this is to defend the risk models of investment banks. They clearly were off in a specific way. The underlying premise appeared to be that a nationwide collapse in housing prices was very unlikely, given the local nature of real estate prices, and the securities were geographically diversified. I actually agree with this premise; but one has to account for low probability correlated risk, even if that risk is that the Fed will screw up. Still, it would be worth it to come to a broad realization that the more fundamental problem was policy error.)
V.
Most of the latter half of the book discusses the specific broader economic effects of our implicit policy to wall off centers of growth and opportunity from a majority of the population, and to also chase away the people that are already there to begin with. This part is not particularly original, and is likely well-trodden ground for readers of this blog. Nonetheless, it’s important to shout from the rooftops - rinse, repeat - until the truth becomes unavoidable. If we are going to develop a collective narrative about housing, it might as well have the virtue of being correct.
“Closed access housing has become such a high barrier to the free movement of capital to its most valued use...that the US population appears to be fleeing prosperity. The effect this is having on our national standard of living - especially for households with lower incomes - must be significant… [Data] shows that rising income is associated with migration away from a city. We should surely expect migration patterns in a liberal society to operate in the opposite direction… But through local housing constraints, we have created an economy that is shaped by exclusion… income is determined by access, and households are pushed away from affluence. Exclusionary housing policies are the root cause of wage stagnation, asset inflation, income inequality, and the “bubble” economy.”
Erdmann presents graph afer graph that confirms the story of low-income residents of high income cities fleeing to lower-income areas, leaving them with more to consume net of housing costs. (More to consume, that is, of material goods; one can only assume the consumption of the non-material good of living in an amenity-rich dynamic city is worth something to these people, or they would have left before being squeezed out.) And as the proportion of the population in high-income cities has shrunk relative to those living in lower-income cities, measured inequality has grown.
The upshot is that economic innovation has slowed down as the barriers to highly innovating locales have been maintained; consumer prices are higher than they would otherwise because of what is necessary to compensate landowners in closed access cities; and economic opportunities are unavailable to the poorer among us. All because of short-sighted local housing policy in a handful of places.
(This was written in 2018; it remains to be seen how the post-COVID economy affects these dynamics. It is plausible that these could rapidly reverse independent of change in policy in these areas, as work-from-home and an extension of the housing-related exodus into higher income brackets could shock the supposed network effects of highly innovative cities, causing a cascading reversal of this trend).
He even shows the effect this had had on international trade. Erdmann (smartly) steers clear of discussion of political party dynamics, but it cannot be a coincidence that the era of the housing boom-and-bust, with different dynamics playing out in different geographies, just so happened to occur along with the acceleration of a great reorganization of political alliances. While problems in education and health care are not due to real estate costs (though, do you know how much your doctor is paying in rent?!), similar dynamics are at play where, in the words of economist Arnold Kling, we effectively “subsidize demand and restrict supply”, and then are surprised by the outcome. The lessons of housing do generalize.
I’m not saying everything is explained by housing costs; I’m just saying whatever issue you’re taking up, see how much, if any, is explained by this dynamic and then see what you have left.
VI.
The only real negative of Shut Out is the necessary slog through a data-rich jungle of information; many of the data figures are great in navigating the terrain, but they can be difficult to read, and some colorized charts would have been very helpful. I admire the author for forging ahead anyway; part of the problem that made the actual dynamics at play so difficult to recognize in real time was the reliance on aggregated national-level data that told the opposite (and false) story from the one elucidated by the disaggregated local-level data in Shut Out.
The reward for doing so is in the epilogue, where Erdmann sings a beautiful hymn to an open and dynamic economy. He talks about the open access orders that have evolved in the developed world (especially in the Anglosphere), in contrast to the closed access orders that persist in the developing world, where opportunity is limited in economies where access is a political favor to be granted.
The housing markets in the richest cities in the US are becoming more like developing world economies in the way that access is granted by either paying unnaturally large sums (for new renters and owners) or by navigating a thicket of economic rents in the form of taxes, fees, and political favors (for developers). Even worse, with a better developed formal legal and political system, enforcement of this closed order is more effective than what you see in the poorer countries. Navigating the closed access order in Brazil or India may be as simple as paying a petty bribe; in S.F., to build new housing you have to pay large fees and convince power brokers at public meetings, and you might still get shut out of the process (but no refunds!).
In the end, the very future of our economic system may hang in the balance; because of the effects of housing policy, and our inability to properly diagnose the downstream problems, “there is a growing sense that spoils of growth are only shared by a select few. The politically imposed limits on how many people can share in prosperity undermines the communal support for growth.” Erdmann seems pessimistic that this can be solved in these markets, given the entrenched interests, and offers the history of decline in Detroit (for reasons other than housing) as a warning. But the rest of America and the cities that might challenge for economic supremacy may be able to learn from the mistakes of the closed access cities, and elect to build liberally. His advice: “let it rip.”
VII.
In college basketball, former UNLV coach Jerry Tarkanian, an outlaw who skirted the rules of amateurism and who was a constant critic of the institutional framework of his sport, once famously quipped: “The NCAA is so mad at Kentucky it will probably slap another two years probation on Cleveland State.” His point was that the NCAA protected larger cash-cow programs and saved their enforcement firepower for the also-rans.
It was probably never quite true of the NCAA, but if the ghost of Tarkanian came back to say “the federal government got so mad at San Francisco for not building enough housing that they engineered a recession and forced Ohio homeowners to lose their homes”, I’d say it was the most accurate (and fair) description of the late 2000’s that had been uttered.
Indeed, if the engineered housing bust and recession had somehow solved the housing supply crisis in superstar cities, I suppose it would be worthwhile to debate whether it was worth it, even if there were other better ways to solve the problem. But it solved nothing. More than 10 years on, the cost of housing and rents in these locales is as high as they were during the boom years (temporarily falling rents due to COVID notwithstanding), while home prices in the rest of America still haven’t recovered and potentially willing and able homeowners still can’t qualify for a mortgage.
But I doubt any of these problems can be solved without a reckoning of how most everyone got the narrative wrong, and how the imposed solutions likely made things worse in the long run. Hopefully, Shut Out is the beginning of that reckoning.