The Connectedness of Our Housing Ecosystem

Many American cities are dealing with a housing affordability crisis. And it's not one that will yield to simple solutions, despite the abundance of simple rhetoric out there, like"There aren't enough affordable homes. So we should build more affordable homes." With what money? The resources that most local governments have to attack the problem in this direct, blunt-force way are nowhere near the scope of the challenge.

But that doesn’t mean it’s hopeless. The key is in understanding what kind of problem housing is.

Housing markets—and indeed, cities in general—are more like complex ecosystems than simple, mechanical systems. They're comprised of the individual decisions of thousands or millions of people—decisions which are interrelated in ways that aren’t always apparent.

A fascinating bit of research which helps drive this point home recently came to my attention. Economist Evan Mast of the Upjohn Institute analyzed the effect of building new market-rate housing in 12 cities on the supply of affordable homes elsewhere in the same cities. This is a topic that's been studied dozens of times in different ways, but Mast's approach is very novel: he actually traces and analyzes the chain reactions that are set off when families move out of homes and other families move into those homes.

The result is a glimpse into just a small part of what we mean when we talk about cities as complex systems. Pretty much any action you can take involving the built environment—building a new apartment complex, opening a grocery store, expanding a highway, and so forth—sets off an intricate web of causes and effects that ripple down the line in ways that are difficult to perfectly trace. It's only through an intense amount of effort and simplification that Mast is able to trace even one aspect of them.

And this has major implications for how we need to think about planning for the outcomes we'd like to see in our cities—such as a place where everyone who needs a place to live can find and afford one. We won't achieve this by pulling a policy lever—”Build more affordable homes”—and getting a desired result as a direct, linear consequence. We have to more subtly try to influence incentives and get the system in balance. We have to figure out why the market—by which I mean those decentralized decisions of thousands or millions of people—isn’t getting people into homes they can afford. And most important, where and how public policy is playing a role in stopping it from doing so.

The engineer tinkers with what’s broken in a machine. The conservation biologist, on the other hand, tries to help a system get into a self-sustaining equilibrium.

The planner's job needs to have more in common with conservation biology than with the tinkering of an engineer.

Mast's Finding: New Expensive Housing Frees Up (Some) Older, Cheaper Housing

The Mast paper is titled, "The Effect of New Luxury Housing on Regional Housing Affordability." Keep in mind that "luxury" is arguably a misnomer here; what Mast means (and he says this) is market-rate housing, built without any subsidy and rented or sold to inhabitants near the high end of the market.

“It’s not at all intuitive to many people why “luxury” housing would do anything at all to help poor people find homes they can afford. The reason it does is these cause-and-effect migration chains. We’re all more connected than we think.”

Mast used a commercial database of exact address-change data to construct migration chains beginning with residents of 802 new, high-end buildings. He categorized neighborhoods by income decile (i.e. the bottom 10%, the 10th to 20th percentile, and so forth) and then analyzed chains of relocation. So, for example, when a housing unit opens up in the 8th decile (70th-80th percentile) because someone moves out of it, Mast looks at the household moving in: what kind of neighborhood did they move from? And what about the place they vacated? Who moved in there? And so forth.

There's a lot more detail you can read about in Mast's paper—for example, he accounts for the ways chains can be broken, such as new household formation (for example, a 20-something moving out of their parents' place to get their own apartment) or an owner taking a unit off the market. I won't summarize it all here. The key finding is that these chains work their way down the income scale and, after a few steps, many of them reach even the poorest neighborhoods.

Mast created a theoretical model, calibrated based on the actual address chains he analyzed, and found the following estimation:

Building 100 new luxury units leads 65 and 34 people to move out of below-median and bottom-quintile income neighborhoods, respectively, reducing demand and loosening the housing market in such areas. These results suggest that increasing housing supply improves housing affordability in the short run.

This is roughly in line with previous findings, like a UC Berkeley study by Karen Chapple and Miriam Zuk that found that two new market-rate homes were roughly equivalent to one affordable home in terms of reducing the displacement of low-income people.

What Mast does, however, is give us a glimpse into the actual mechanism by which that would be true. It's not at all intuitive to many people why "luxury" housing would do anything at all to help poor people find homes they can afford. The reason it does is these cause-and-effect migration chains. We're all more connected than we think. 

The "Six Degrees of Kevin Bacon" City

Our friend Joe Cortright at the think tank City Observatory (whose work we frequently publish on Strong Towns) describes Mast's findings in terms of the party game Six Degrees of Kevin Bacon, in which players try to connect any show-business figure to the prolific actor Kevin Bacon through the shortest possible chain of people who have appeared on screen with each other:

So for example, John Turturro’s “Bacon number” is 2:  he and Julianne Moore were both in the cast of The Big Lebowski, and Moore in turn played opposite Bacon in Crazy Stupid Love. 

The Kevin Bacon game, much like the Six Degrees of Separation theory that is its namesake, suggests the world is smaller than we think, and we're all more connected than we think. Mast's research suggests that is true of our homes. Within a few steps of a chain in which someone moves out of a home and someone else moves in, we can get from a very rich neighborhood to a poor one.

This challenges the idea that housing is rigidly separated into "submarkets" and that what we build over here has no effect on the circumstances of those people over there.

An Ecology of Causes and Effects

Lurking below the surface of this study is a much more profound implication about cities, and a point that is crucial to the Strong Towns approach: cities are complex adaptive systems. Every part is connected to every other part in a web of causes and effects far too intricate to trace.

Mast's housing chains are in fact a deep simplification of this web. For example, he traces moves based on the average income level of a neighborhood, because he does not have detailed data about who is actually moving. Perhaps only the richest people in poor neighborhoods are able to take advantage of a newfound opportunity to move out. And what constitutes a poor or rich neighborhood is itself a moving target, determined in part by—you guessed it—who moves out and who moves in. And what about forces that affect the migration rate? (Mast assumes that every household that migrates to a metropolitan area from outside would have done so regardless of whether any new housing was built... but at some level, this is certainly not true.) These are not flaws in the study; they are limitations in any one study's ability to grapple with complexity.

If Mast's results hold, a new high-end apartment building with 100 homes will result, on average, in somewhere around 30 or 40 affordable homes becoming available to low-income people, somewhere. We don't know where. We don't know who will occupy them. We don't know how that will alter the composition of neighborhoods—who lives in them, who do the local businesses cater to, who has social support networks there.

This has implications for how we need to practice planning, and Mast gets into a couple of these, but I'm going to extrapolate beyond Mast and his work here, and discuss two broad implications.

“All of these counterproductive policies have something in common: they’re rooted in linear thinking, rather than ecological thinking. They all say, essentially, “We have a problem with not enough (Blank). So let’s spend some money and make more (Blank).” The problem is that when you do that, you’re changing people’s incentives and thus their behavior... and what they do differently may erase some or all of the gains you just spent a lot of public money on.”

1. Policies that are intended to do good can end up doing net harm. The specific example of this which Mast does invoke, in the realm of housing policy, is Inclusionary Zoning. This is a policy that requires that developers set aside a certain percentage of the homes in a new residential development for households below a certain income, at rents that will be affordable to them. (Sometimes the developer gets an incentive for doing so, such as additional density or height; other times they don't.)

The problem with Inclusionary Zoning is that, by imposing additional costs on the developers of new housing (requiring them to build units that must be rented or sold at a loss), it may decrease the amount of such new development that occurs. If it does, according to Mast's findings, it is directly eliminating some of those migration chains that would have caused homes in low-income neighborhoods to be freed up for new occupants. 

Examples abound, beyond housing, of policies that are well-intended but counterproductive. Widen a highway to relieve congestion? Traffic shifts to that highway, or people take trips they would not have taken, or people buy homes farther from their jobs than they would otherwise have lived, and pretty soon it's congested again anyway.

All of these counterproductive policies have something in common: they're rooted in linear thinking, rather than ecological thinking. They all say, essentially, "We have a problem with not enough (Blank). So let's spend some money and make more (Blank)."

The problem is that when you do that, you're changing people's incentives and thus their behavior... and what they do differently may erase some or all of the gains you just spent a lot of public money on.

2. Practice via negativa instead. The via negativa is an approach recommended by Nassim Taleb for working with complex systems: work via subtraction, not addition. Instead of adding distortions to the incentives that people face, take away existing interventions that are distorting things.

In the case of housing markets, this means letting feedback mechanisms work. Where you have a productive place that people want to be, it should be possible for that place to naturally "thicken up" over time—not in large leaps that risk killing the golden goose that made it a successful place, but in small increments. This means letting people build more homes—yes, even if some are "luxury" homes. It means letting them build a diversity of home styles and sizes, so that the emergent wisdom of the crowd can shape the evolution of a place.

Mast’s migration chains suggest that this approach may even do more to produce broad housing affordability than explicitly mandating capital-A Affordable housing, and micromanaging where it is built. The housing market is a complex system, and its overall outcomes—who finds a home at a price they can afford in a place they want to be—are going to largely be shaped by an ecology of causes and effects that defies micromanagement or simplistic understandings of cause and effect.

---