The Vicious Cycle of American Housing

For a century, housing has been stuck in a vicious cycle of layering bad policy atop bad policy. It is imperative that we end it.

The Vicious Cycle of American Housing

The last few decades of American housing have passed through several phases of a vicious policy cycle. There may be nothing more urgent than ending it.

A polity can find itself in a vicious cycle where bad public policy leads to bad outcomes. Instead of retracting that policy, additional policies are layered atop it, each with their own unintended consequences, which, in turn, attract more layers of new policies.

This is common with issues like immigration, drugs, and sex. In those arenas, one sign of ineffective policy is the rise of black markets. You might say that the increasingly common tent encampments in our major cities are our housing black markets.

Policy Cycle 1: Closed Access Cities (mislabeled the Superstar Cities)

The policy vicious cycle began in the early twentieth century when major American cities started to experiment with local zoning ordinances. The vicious cycle was codified in the 1926 Supreme Court decision, Euclid v. Ambler, where the court called apartment buildings parasitical and opened up a century where city-building as it had necessarily occurred in the past could be outlawed as a nuisance.

Some cities which had developed before zoning have declared themselves illegal. The New York City planning department, ironically, resides in a building that the department would be required to reject if it were proposed today and that is also, officially, a National Landmark.

Closed Access outcomes

By the end of the twentieth century, the New York City, Los Angeles, Boston, and San Francisco metropolitan areas had encased themselves so securely within arbitrary rules against city-building that they lost the ability to grow. I refer to these as the Closed Access cities. From 1994 through 1999, annual permits for new homes had fallen below 3 units per thousand residents in the Closed Access cities.

About 6 units per thousand residents are required to maintain low, normal population growth. Among other cities, metropolitan areas like Detroit and Chicago were permitting about 4 units per thousand residents at the time. Successful, growing cities permit well more than 10. The Closed Access cities had become extreme low outliers in housing production. Their ability to house residents had become unsustainably low.

Residents of the cities that lack adequate housing must engage in a game of housing “musical chairs”. When there aren’t enough new homes, residents must bid for existing homes. For rich residents, housing is mostly a luxury. So when the bidding war pushes prices up, rather than pay more, they opt for less. In practice, that means moving to a neighborhood that would have previously housed families who were slightly less rich.

For families with lower incomes, housing is less of a luxury and more of a necessity. At some point, there are no more bedrooms to give up. No longer commute to bear. It becomes a stark choice between paying more, choosing to be regionally displaced, or becoming homeless.

Economists call this filtering. In cities that allow enough new housing, older homes become more affordable. New residents of older neighborhoods typically have lower incomes than the previous residents. Where housing is in short supply, the filtering goes in the other direction. The new residents of older homes tend to have higher incomes than the previous residents. The prices of those homes rise.

So, when housing is in short supply, rents and prices rise the most in the poorest neighborhoods. Their residents have the fewest options. This leads to a self-selection of remaining families. Families either have high enough incomes to trade down and remain in the city, or they are willing to spend more of their lower incomes to avoid displacement, or they are displaced. When about 1% of a city’s population is forced to leave each year, after a couple decades, the compositional effects of that exodus add up.

New research is starting to highlight the effects of this problem. Researchers have been discovering that the high gross incomes of the Closed Access cities are not high once the cost of housing is accounted for. It is increasingly clear that residents of the Closed Access cities don’t have higher incomes because those cities are “superstars”. Those cities have instead failed their poorest residents. The typical college educated resident of the housing deprived cities is just breaking even compared to similar workers in other cities. And, workers with less education are worse off than workers in other cities. The difference in incomes between cities is mostly due to housing displacement. The poorest residents have to make a hard choice. They either choose poverty over displacement by paying the high cost of housing or they choose displacement, which raises the average incomes of those who remain.

Misunderstanding Closed Access outcomes

In the past, successful cities didn’t become such outliers in average incomes. They mostly grew larger. Every new engineer that moved to Detroit in the early twentieth century was joined by teachers, police officers, waitresses, clerks, etc. The average Detroit income was only moderately higher than other cities. Mostly, it grew.

Today, the cities with high incomes are the slowest-growing cities in the country. Their rates of net domestic migration are the worst in the country. An anthropologist looking at only their population and migration trends would assume there had been a famine or a natural disaster. It is their outmigration that is unusual.

Are those cities popular? How would we know? There aren’t actually hordes of newcomers moving in. We are left only to presume that phantom newcomers would like to move in. Locals only have the impression that there are many newcomers because, without more housing, each newcomer brings problems for locals instead of opportunities. From 1995 to 2022, the Los Angeles metropolitan area grew by a total of 10%. Dallas-Fort Worth grew by 77%. The US by 25%. It isn’t remotely clear that demand for living in Los Angeles is higher than Dallas.

Policy Cycle 2: Housing Refugees (mislabeled the Housing Bubble)

As the century turned, these pressures mounted. Each year, migration away from the Closed Access cities increased. By the mid-2000s, net domestic migration reached about 1.5% of the Closed Access population annually.

Housing Refugee outcomes

The cities they moved to (Florida from the northeast, and Arizona, Nevada, and inland California from the California coast) became the emblems of a housing bubble. Eventually, the exodus from the Closed Access cities overwhelmed the ability of growing cities to grow even faster, and home prices in those cities started to rise. Those high prices were the motivation of countless retrospectives from Mian and Sufi’s “House of Debt” to Michael Lewis’ “The Big Short”.

Many outcomes you might think happened, didn’t. During the subprime mortgage boom, the convention says that “anyone with a pulse” could get a mortgage. Yet, homeownership peaked by early 2004, before the subprime lending boom and well before the CDO markets highlighted in the Big Short were in vogue. The income of the median homeowner actually increased between the mid-1990s and the mid-2000s, relative to the median renter. New homeowners were more likely to have a college education and work in the professions. Homeowners weren’t buying bigger or better homes compared to renters. By 2007 lending was becoming risky, but it was families with the highest incomes who were taking on mortgage payments that claimed more than 40% of their incomes, not families unqualified for home ownership.

Misunderstanding Housing Refugee outcomes

The terms of mortgages became riskier but, all considered, the qualifications of new homeowners generally did not. Families were playing housing “musical chairs”. And, “musical chairs” attracts speculators.


The high cost of housing in the poorest neighborhoods of the Closed Access cities led to a popular belief that loose lending fueled the price bubble by allowing poor families to take on untenable mortgages. When I controlled for the regressive effect of inadequate housing supply, I found evidence for a moderate credit boom. Expanded credit access, on average, was associated with an increase in home prices of about 8% from 2002 to 2006, which then reversed. (That is actually a larger economic effect than other research finds.) But, most of the boost in prices associated with credit access was in markets where prices and construction activity were moderate.

The much more important factor raising prices was inadequate housing in the Closed Access cities. Prices remain elevated today in those cities in the same way they were then, with much tighter lending practices. Prices in the cities that took on their wave of housing refugees temporarily spiked and then reversed after the wave of migration dissipated during the Great Recession.

The cities that families were flooding away from and the cities that families were flooding into looked different in subtle but important ways.

Prices generally only appreciated excessively in the poorest neighborhoods of cities that were collectively losing hundreds of thousands of existing residents each year to that faminesque migration. But the temporary rise in prices in the quintessential “bubble” cities where those families fled to was not particularly higher in poorer neighborhoods.

Researchers who did not correctly center Closed Access migration in their study designs simply did not look for those differences.

Construction Boom

Many of the worst policy decisions were based on the idea that a credit bubble had created a glut of housing. In reality, there were three types of cities, none of which fit that description. Construction activity increased in the Closed Access cities and the Contagion cities that took on their migrants. Vacancies were low there. They really needed homes. Prices and construction activity were moderate elsewhere.

On a per capita basis, national housing production was never very high, and the small amount of increased construction had reversed by the end of 2007, even in bubble regions like Nevada and Arizona. The peak of vacancies in 2009 and 2010 happened well after the collapse in construction had more than reversed any additional building in any city. Vacancies were due to a broken economy and the arrest of decades-old migration trends, not a glut of homes. The record is clear.

Policy Cycle 3: Lending Prohibition (mislabeled the Housing Supply Glut)

If prices in 2006 were mostly high because of supply constraints and there wasn’t a supply glut, then why the collapse?

Lending Prohibition outcomes

The origination of the fragile mortgages in the new private subprime and Alt-A mortgage markets dried up by early 2007. Then, the pendulum swung too far.

Data from the New York Federal Reserve and from Fannie Mae concur. The average credit score of new borrowers remained relatively stable from the 1990s through 2007. In 2007, the average price of homes getting new Fannie Mae mortgages and the average price of homes with existing mortgages were both near $250,000. Prices had appreciated significantly. In 2000, the average price of both had been about $150,000. But, while the values of homes had increased, Fannie Mae was issuing mortgages to the same borrowers in the same houses that they had been for years. Average prices were rising because of the urban housing shortage.

Then, by 2009, the average credit score of new borrowers permanently rose by about 40 points. Home values across the country had collapsed, and homes with Fannie Mae mortgages were no different. By 2009, the price of the average home with existing Fannie Mae mortgages was down to $204,000. But homes with new Fannie Mae mortgages in 2009 averaged $327,000. Between 2007 and 2009, lending to the bottom half of the American housing market imploded and never returned. The federal agencies abandoned their existing customer base and regulators forced even stricter standards on banks.

Misunderstanding Prohibition outcomes

This was not a reversal of a bubble. It wasn’t inevitable. It wasn’t necessary. And I would argue that it was responsible for the entirety of the $5 trillion net loss in real estate wealth between 2007 and 2012. Atlanta provides a keen view into what happened. Atlanta was a metropolitan area that didn’t have a building boom or a price bubble. Real rents declined by about 15% in Atlanta between 2002 and 2006. The moderate American building boom had been a boon to Atlanta’s renters.

In neighborhoods with incomes around $150,000 or higher, Atlanta homes typically sold for about 3 times the average neighborhood income, with surprisingly little variation through boom and bust, from 2002 to 2019.

In the typical neighborhood with an average income of about $50,000, from 2002 to 2007, homes sold for about 4 times income. Again, with surprisingly little variation. Then, when lending dried up, home prices in Atlanta neighborhoods with $50,000 incomes collapsed down to 2 times the typical tenant’s income by the end of 2011. Cut in half.

Home prices before 2008 across Atlanta were typical of many cities. It is common for the price/income ratio of homes to be slightly higher in poorer neighborhoods than in richer neighborhoods, even in cities with ample housing, like Atlanta in 2007. In fact, it would be quite odd for it to be the other way.

By 2011, in Atlanta, it was the other way. It would have taken the typical working-class resident of Atlanta fewer years to pay for the home they lived in than it would have taken the typical rich resident of Atlanta. This remarkable reversal of the norm could only be achieved by making mortgages to those families effectively illegal.

Add up all those trends in city after city, and the tightened lending standards after 2007 pulled $5 trillion out of the wallets of American homeowners in direct proportion to how little they had to begin with—a deeply regressive moral panic.

Policy Cycle 4: The Rent Crisis (mislabeled the Build-to-Rent boom)

The suffocation of entry level mortgage lending meant the suffocation of entry level construction. Before 2006, more than half a million new homes were regularly sold each year for less than $200,000. By 2011, barely 100,000 were sold at that price point, and that market has never rebounded.

Rent Crisis outcomes

When construction dried up, rents increased. Construction dried up the most where incomes were low and homes had lost the most value. In Atlanta, construction dropped by 90% even though it hadn’t risen during the subprime boom.

So, rents subsequently increased the most where home prices had fallen the most. Rents from 2015 to 2021 rose about 40% more in the poorest neighborhoods that had the deepest price drops than they did in the richest neighborhoods where prices had been more stable after 2007.

Now, the entire country was building at a rate below 3 units per thousand residents. The entire country was playing housing musical chairs. Since this cycle of housing deprivation was caused by national mortgage suppression rather than local land use policies, costs increased in poor neighborhoods in every city, and they increased the most in the poorest cities.

Construction of apartments had been strong right up until the September 2008 crisis, and then recovered quickly back to pre-crisis activity by 2012. But all cities have made multi-family and infill housing too difficult. The United States needs well over a million new homes a year. It has been decades since more than a few hundred thousand apartments have been approved nationally in a given year. And the time it takes to complete an apartment building has doubled since the 1990s. So, apartments have not been able to fill the gap to produce a sustainable amount of housing since 2008.

When prices collapsed, investors and corporations started to buy single-family homes. A new industry of large corporate landlords developed. Finally, rents have increased enough that existing homes sell for as much as new homes. Onerous post-crisis mortgage regulations remain in place. The tenants for these homes generally cannot buy them. So corporate landlords order the new homes.

Misunderstanding Rent Crisis outcomes

Naïve observers blame the corporate landlords for the rising prices and rents. And, so, now, at several levels of government, bills are being proposed to exclude these buyers from the market. If we can keep Wall Street money out of local housing markets, prices will go back down. This is true. Rents, then, would continue to go up.

This is worrisome. This is the last round of the vicious cycle.

Let’s review our national devolution. In turns, we have restricted:

  1. All housing in a few key metro areas.
  2. Multi-family or infill housing in every major metro area.
  3. Single-family mortgaged entry level homes in every major metro area. And now, potentially,
  4. New single-family rental homes.

That is all the forms of housing, short of tents and underpasses. We are one step away from legislating mass homelessness.

Policy Cycle 5: The Homeless Crisis (mislabeled the Drug and Mental Health crisis)

The basic pattern will repeat, and in the cities with the worst housing conditions, we already have a preview of cycle 5.

Homeless Crisis

Where housing is in shortest supply, tent cities are accumulating in our parks and on our street corners. Of course, the residents that end up homeless aren’t middle managers and engineers. They are the most vulnerable residents. Where homes are ample, mentally ill residents mostly have a home to sleep in. The rate of homelessness across the US is correlated with low housing vacancies and high rents. It isn’t correlated with mental illness or drug use.

Misunderstanding Homeless Crisis outcomes

Since many homeless residents have personal problems, it seems obvious that personal problems are the reason for their condition. And they frequently are! But the reason a lack of housing is part of that poor condition is poor local housing supply.

This is just one more case of mistaking the side effects for the cause. The alternate cause of every phase’s problems has been logically unassailable. The Closed Access cities are popular. Loose lending can push up prices. There was a building boom. It was followed by vacant units and a foreclosure crisis. Corporations do outbid families for homes and they have raised rents.

We can continue down the same path by blaming mental health and drugs for homelessness. Bulldoze the tent encampments. Oppose the new housing support center because it’s too dangerous to have those people living so near the neighborhood.

Each of these steps has been a test of our liberal birthright. And we have collectively failed, step after step, decade after decade.

I tremble to imagine what Policy Cycle 6 will be.

Ending this

In addition to the cycles of failed housing policies, Americans have also sought to blame various sources of demand for costly housing: short-term rentals, wealthy foreigners, immigrants from the south, low interest rates. A conventional description of the Great Recession is that the housing bubble fooled Americans into thinking we were richer than we really were. You could say that the Great Recession was our attempt to make ourselves poor enough to fit into our inadequate stock of housing. And it briefly succeeded. Maybe the most popular idea of the past 20 years among the financial layperson is “Raise interest rates until the bubble pops.”

Before 2008, even at the bottom of our deepest recessions, America never produced fewer new homes per capita than we produce today. What does it say of America if this is now too much housing demand for us to handle?

It clearly should not be. And to recognize that is to rediscover an optimism for what we are capable of if we simply allow it.

Illiberal assumptions have led to illiberal policy choices, which have led to illiberal conclusions, over and over again. Our very perceptions, our self-identity as Americans, our mutual trust, are mired in a fog of confusion.

Home prices in the Los Angeles metro area are at record highs. Yet, its population has declined by 3% since 2017. How high do interest rates need to rise? How much unemployment do we need to create? How many immigrants do we need to deport? How many Airbnb’s do we need to shut down? How many tents do we need to bulldoze? How illiberal are we willing to be to push Los Angeles’ population down far enough to fit into its pitiful housing? Or will we choose, at this late date, to pay forward the growing, aspiring nation we inherited?

Featured image is Sparks, Nevada, by Ken Lund