The Housing Market Never Recovered from the Great Recession

(Credit where due: This post owes an intellectual debt to a Twitter thread by Payton Chung.)

If everything you thought you knew about housing turns out to be wrong, it's probably not your fault. It's actually just that we've just had a really weird, unprecedented, decade.

You wouldn't know it from the tenor of most media coverage of housing. The dominant themes of national reporting, at least since about 2014, have been of a return to normalcy. Prices are back up above their previous peak, construction is booming again, and, we're told, buyers are once again opting in droves for the Sunbelt and the suburbs, just as they did in the pre-recession world. (To that last point, though: there's more to it than meets the eye.)

Of course, the run-up to the Great Recession was far from normal in its own right. (If you've never seen the roller-coaster animation of the Case-Shiller index of inflation-adjusted home prices over the past century, check it out. Specifically, check out what happens toward the end.)

But it also turns out that we've never really returned to any sort of pre-recession status quo since. Three key facts illustrate this:

1. Construction levels never recovered.

Despite the popular perception that big cities boomed in the 2010s, it was actually the weakest decade for housing starts in a very long time. Freddie Mac has some numbers here. Here’s a graph that illustrates the cliff we fell off after 2007. Keep in mind these are absolute numbers, not even adjusted for population growth over the past half century:

 
 

How do we reconcile this with the evidence in front of many people’s eyes: construction cranes galore in hot neighborhoods of many U.S. cities? The simplified answer is that the cranes are only in hot neighborhoods, but they’re heavily concentrated in those few places, so the high visibility of change makes a strong impression. Outside of those places, your city probably isn’t changing as fast as you think.

Call it the Trickle or Fire Hose phenomenon: change in pretty much every American metro area is concentrated into only a small minority of neighborhoods. Most areas receive a mere trickle of reinvestment and may see almost no new development for decades (often while maintenance is deferred on buildings and the public realm and things fall into decline). A much smaller number of trendy ZIP codes are the target of a fire-hose blast of outside money which buys up, transforms and redevelops properties until the place is unrecognizable.

This trend has only been accentuated in the 2010s, as development has become dominated more than ever by a small oligopoly of deep-pocketed companies—thanks in part to regulatory and financial institutions that make it punitively difficult for small-scale builders to get into the game.

2. Homeownership levels plummeted.

The proportion of U.S. households that were owner-occupied hit an all-time high of about 69 percent at the peak of the early-2000s boom. This was fueled by a dramatic expansion of easy credit, and came to an abrupt end with the housing market collapse and subsequent rash of foreclosures.

As this data from the St. Louis Fed illustrates, it took until 2016—well into the resurgence of housing prices themselves—for the homeownership rate to bottom out at 63 percent.

 
 

A 6 percentage point difference might not seem like a lot, but in absolute numbers, the country added over 10 million renter households in the period from 2005 to 2016, during a period of historically weak housing construction. This had a really important consequence, which is our third and final trend:

3. Filtering reversed direction after 2010: it's now mostly upward.

This one comes from a new study from the research arm of the National Multifamily Housing Council (NMHC), conducted by demographers Dowell Myers, Ph.D. and JungHo Park, Ph.D. from the University of Southern California.

The study, which you can download here, examines the filtering process, something we've covered at Strong Towns before. Filtering is the phenomenon by which homes built for people in a particular income strata become occupied, over time, by people either poorer (filtering down) or richer (filtering up) than their original inhabitants. There is no question that this occurs—we're all familiar with former mansions that have been subdivided into modest apartments, or with historic bungalows in once-poor-but-gentrifying areas that have been renovated and flipped for a pretty penny.

But just how fast filtering happens, how pervasively, and in which direction(s) are the subject of much academic research.

Conventional wisdom, including what I was taught in my graduate-school urban economics courses, is that housing usually filters down over time. It makes sense: an apartment's size, features, and aesthetic may become dated, and what was an up-and-coming neighborhood may go into decline or at least be less trendy.

Turns out this can be measured, and doing so results in easily the most startling graph in the new NMHC report:

 
 

Here’s what this is telling us. Between the years of 1980 and 2010, apartments built in every single decade since the 1960s tended to filter down—that is, the share of them occupied by low-income people increased over time.

After 2010? Apartments built in every single one of those decades began to filter up instead: they became occupied, on average, by higher-income people. This has led to a huge, systematic decline in the supply of apartments for low-income Americans, something you can see broken down by metro area here (the red bars reflect the post-Great-Recession trend):

 
 

Using a regression analysis, the NMHC study finds that the decline in homeownership especially among young adults directly contributed to reducing the supply of affordable apartments, by introducing a much greater level of competition for the available ones:

[A] one percentage point increase (decrease) in the age 25-34 homeownership rate is associated with a 0.521 percentage point increase (decrease) in low-income occupancy of apartment units. This finding spotlights the interconnection between rental and owner markets. Specifically, this implies that an increase in the homeownership rate among young adults eases rental competition and opens greater opportunities for low-income renters. Instead, after the Great Recession, homeownership fell ten percentage points for young adults, and filtering was reduced proportionally.

The Careening Car

What we’re looking at is a housing market that is unable to easily adapt to changing needs and economic circumstances. Construction is at historically low levels and not consistently happening in the places where there’s high demand to live. The result is a set of polarized, unpleasant outcomes: affordability crises in some regions, cities, and neighborhoods; decline and economic dislocation in others.

For my entire adult life and that of my whole generation (I’m a 35-year old Millennial), the U.S. housing market has been a lot like a car skidding on ice, careening in one direction and then overcorrecting in another. With the current coronavirus-induced economic downturn, there’s good reason to think this will continue.

It doesn’t have to, though. A Strong Towns approach to development would work to systematically eliminate some of the sources of fragility and lack of adaptability in the housing supply—the things that cause the “trickle or fire hose” problem, the things that make the system prone to overcorrecting itself repeatedly.

Want to be part of that conversation? Here’s some further reading.

Herriges, Top StoryDaniel Herrigeshousing market, housing affordability, filtering