The Urbanist Musings of Pete Saunders

How Rust Belt Recovery Looks Different From Everyone Else

Block for sale, West Village, Detroit.  Source:

There are people who believe that population loss in Rust Belt cities, a long-term phenomenon, should be a great concern.  There are those who don’t.  More and more, evidence suggests that population decline in the Rust Belt may be structural, even inevitable.  However, economic decline is not.

Daniel Kay Hertz of City Notes sits on the former side of this argument.  In a recent blog post, he surmises that restrictive zoning policies in Chicago, where the prevalence of single-family home zoning is putting a lid on construction activity in the neighborhoods with the strongest demand, is keeping Chicago from reaching its growth potential:

“For a long time, most accounts of Chicago’s lagging population have focused on parts of the South and West Sides where many residents, largely African-American, have decided to decamp for the suburbs or the South in search of better schools, less crime, and more jobs.

But the under-appreciated flip side of population loss in those parts of the city is that places that ought to be growing like gangbusters are stagnant, often sitting 25% to 50% below their peak populations. Lakeview, for example, was once home to 124,000 people; its population is now 94,000. North Center is down from nearly 49,000 to under 32,000. West Town, which includes Wicker Park and Bucktown, has fallen from 187,000 to 81,000.

For most of Chicago’s history, when a neighborhood became more popular, builders created more housing, turning houses into three-flats, and three-flats into courtyard buildings. In a few really high-demand areas, like right along the lakefront or near downtown, they might even have built highrises.

But for the last several decades, increasingly strict zoning laws have outlawed this kind of gradual build-up. Instead, Chicago’s laws allow a massive boom in parts of downtown – mostly where there weren’t enough white-collar residents to complain – while putting a tight lid on the neighborhoods.

Since replacing a couple two-flats with a courtyard building is now illegal, developers make money by tearing down an old two-flat and building a luxury two-flat in its place. Or they build a mansion, and the neighborhood actually loses a housing unit. As a result, as a neighborhood becomes more attractive, the city encourages fewer people to live there.”

But there are other urbanists who say that, no matter what Rust Belt cities did, they ran into a demographic buzzsaw that was unavoidable, and that loss was inevitable.  St. Louis blogger Alex Ihnen of drove home this point back in September, demonstrating how two of the most significant American demographic trends of the 20th century impacted his hometown:

American families have changed dramatically over the past half century. The average household size in St. Louis in 1950 was 3.1 and in 2010, 2.2. With every other factor held constant, the decrease in city population would have been 248K or 29%. This means that with the same number of homes, the same number of apartments, and the same number of families as resided in the city in 1950, the decrease in average household size could account for 46% of the city’s population loss.

American homes have changed dramatically over the past half century. The average size of a new single-family home in 1950 was 983sf, and in 2010, 2,438. While this number overstates changes in a long urbanized area, there has certainly been a large increase in the average size of the single-family home in the city. This could be new construction, but is also the common two-family to one-family and four-family to two-family renovations.

If nothing else had changed, and families had simply become smaller, and lived in bigger homes, the city would have lost hundreds of thousands of residents. What this really points to is that even if the city, or select neighborhoods reach zero vacancy – zero vacant lots, zero vacant homes – at today’s density, the residential population will not return to 1950, or even 1970 levels.

Planner and blogger Jason Segedy of Akron, OH, who blogs at Notes from the Underground,  picked up on this point last November.  In his piece, he further drove home the significance of the demographic trends on built-out cities, and the advantage that suburban areas had in overcoming it:

Think about what was going on around 1960. Take an older, industrial city, like Cleveland, Detroit, or St. Louis, that is “built out” – that is, nearly all of the residential property in the city is built at the maximum development density that it is zoned for, and the city has very little, if any, available land upon which to build new housing.
At the same time you have several extremely consequential demographic trends occurring all at the same time:

Rising divorce ratesRising age of first marriageRising life expectancyDeclining birth rates

These trends all began to emerge in the late 1950s, took deep root in the late 1960s, and continued virtually unabated up through the late-1990s.

So, set aside any preconceptions you have about the disadvantages of city living versus suburban living.

Forget, for a minute, the usual suspects in urban decline, such as “white flight”, larger suburban houses and yards, highway construction, increasing automobile usage, crime, declining schools, etc.

Focus instead on the profound social trends which occurred between 1960 and 1990. Those involving:

Marriage (divorce and delayed marriage, resulting in more singles)Health care (increased longevity, resulting in more widows and widowers)Reproduction (the development of “The Pill” in 1960, and legalized abortion in 1973, resulting in less children).

The end result of these trends? A much smaller average household size.

So, if these older cities were unable to build more housing units, they were going to shrink to a significant degree, regardless of the crime rate, or school quality, or “white flight”, or any of the other common explanations for urban population loss.

Similar social trends were also occurring in suburban areas, but one mitigating factor was that these areas were being settled disproportionately by families with children.

Suburban areas also had lots of room to build new housing, while the core cities did not. In a developing suburban area, it did not matter if your average household size shrank a bit between 1950 and 2010. If you added an additional 10,000 housing units in that same time period, your community was going to grow – a lot.

There’s no doubt in my mind that the story of 20th century urban decline in America is the story of a profound shift in demographics and living preferences that adversely affected the places that could do little to adjust to the changes — built-out cities.  Once the demographic and living preference changes took hold, the declining narrative settled in for built-out cities, with the most devastating narrative reserved for Rust Belt cities — racial discord, economic decline, poor schools, rising crime.  This became a negative feedback loop that fueled suburban growth and Sun Belt growth, even to this day.

In recent years, however, starting in the ’90s and gaining steam after the Great Recession, cities began seeing a shift in their demographics that, while not always improving their total numbers, was indeed improving their prospects.  East Coast cities like New York, Boston and Washington began leveraging the great numbers of young and talented professionals attracted to them for their benefit.  Similarly, the booming tech economies of Silicon Valley and Seattle began attracting young talent as well.  What was once a structural matter impacting these cities has been largely reversed — as demographics and living preferences shift once again.

Rust Belt cities may just now be cashing in on this emerging trend.  Here’s one (totally cherry-picked) example of what may be happening within our nation’s metro areas.

Most people seem to have an intuitive sense of the broad economic trajectories of Atlanta and Chicago.  Atlanta has remained a strong performer among Sun Belt cities, continuing to attract migrants from around the country.  Meanwhile, Chicago has made a comeback economically that separates it from its Rust Belt brethren, but still faces challenges related to crime and economic inequality.  A couple of charts that illustrate this:

Atlanta significantly outperforms Chicago in population growth in the post-Great Recession era (and for decades before that, actually).  Recent job growth data also conforms with this narrative:

But there is one measure where Chicago substantially outperforms Atlanta — in gross domestic product per capita:

Slow-growth Chicago economically outperforms faster-growth Atlanta?  How can that be?

I can hardly provide a definitive answer, but I’ll offer one thought.  The people who are still relocating to Atlanta are the ones drawn to the demographics and living preferences that characterize the migrants of the last half-century: attractive climate, lower cost of living, housing affordability, among others.  In many respects they are locked out or they’ve given up on the economic possibility of their former homes.  But that doesn’t mean that their former homes aren’t growing.  In a changing economic environment that is rewarding metro areas skilled at attracting talent, metros like Atlanta may be losing out on the productivity scale.  Chicago’s mix of Fortune 500 headquarters and top-tier universities, as well as desirable walkable urban neighborhoods, might be paying off.  As for Atlanta, it may be adding a greater number of jobs than Chicago, but it could be that the jobs the region is adding are lower wage and don’t contribute as much to the productivity of the regional economy.

Could improved economic productivity and a shift in living preferences lead to a change for the Rust Belt?  I think the next 10-15 years will tell.  But from my perspective, the Rust Belt stands a better chance of recovery and renewal today than at any time in my lifetime.

4 Responses to “How Rust Belt Recovery Looks Different From Everyone Else”

  1. Alon

    I don't think the Atlanta per capita income decline is about migration. I don't have statistics either way, but back when Telestrian existed, I used it to look up data for Texas, New York, and California. Migration dragged Texas's per capita income down and boosted California's, but in both cases it was by a very small amount, I believe 1% each, over the decade. New York didn't even have that 1% effect – in other words, the fast growth in the state's per capita income last decade was entirely real, not just an artifact of poorer people moving to the Sunbelt. Likewise, Texas's per capita income growth (which was actually decent) was not materially dragged down by poor people moving in from California, New York, and Mexico.So, without having seen numbers for that part of the South, my guess is still that the decline in Atlanta's real per capita income since 2000 is real, and not just an artifact of poor people moving in from the North.The issue here is that truly poor people aren't migrating to the Sunbelt, because, at least in New York, they have public housing or rent stabilization, and between that and not needing to own a car, their cost of living is actually lower than in the Sunbelt. The Sunbelt is cheaper to live in if you're middle-class or rich, but if you're poor, you're better off taking the subway in an older city. The result is that the Rust Belt-to-Sunbelt migrants, while still poorer than the average residents of both, are not too much poorer – they tend to be working- and middle-class and not poor (who gain nothing from moving) or rich (who prioritize access to high-paying jobs and also enjoy high living standards in rich urban neighborhoods).Another aspect of the Rust Belt that's worth mentioning: the secondary cities of the Northeast have done well since 2000: Buffalo, Rochester, Syracuse, Albany, Providence, New Haven, Hartford, Springfield, Pittsburgh. I believe that Worcester is the only one not to have had substantially higher per capita income growth than the US average; in some of those cities, such as Providence, the growth was twice as high. In some of them, the central city still sucks (hi, Providence ex-East Side), but the metro area is doing well. The Midwestern metro areas have not seen any such growth. It's worth investigating what conditions have existed in the Northeast in the last 20 years and whether they exist in the Midwest now, encouraging such growth in Milwaukee, Indianapolis, Cincinnati, St. Louis, etc.


  2. Anonymous

    In the Pittsburgh metropolitan area, there are more deaths than births. The glut of old people that Pittsburgh had in 2000 wasn't going to live forever. They're all dying now, and it's happening so fast that the birth rate can't keep up. It explains why the Pittsburgh metropolitan area had essentially zero growth according to the 2013 population estimate, even though its net migration was positive by more than 3,000. Basically, net migration has to be more than 3,000 in order for Pittsburgh to break even. Unfortunately, because the 2014 population estimate occurred at the same time as a brief lull in job growth, net migration was positive by less than 1,000, and the Pittsburgh metropolitan area was estimated to have lost a few thousand. Natural decline has become a demographic ball and chain for Pittsburgh, even though more people have been moving in than out for at least eight years now. The good news is, the lull in job growth during 2013 and the first few months of 2014 was a fluke, and job growth has picked back up again, which means that this year's estimate (released next year) will more than likely be much better than the one we're analyzing today.


  3. It's all power politics

    Other demographic trends include:- 71% of married college grads… are married to other college grads- For college-educated women, the average age of first birth (30) has risen along with the average age of marriage (27). Only 12 percent of births by college graduates are to unmarried women.- For women who graduated from high school and went to some college, the average age for first birth is 24 and the average marriage age is 27; 58 percent of first births are to unwed mothers.So the bottom line is, college-educated women are reaping most of the benefits of later marriage: They can enjoy the greater economic security that comes with marrying later, while still being able to have children in the relatively stable context of marriage. Women with lower education levels get a much smaller economic bump for marrying later and are less likely to to be married when they have their first child. rise of Dual Income / Dual Degree Households that do not have children until they are 30+ is driving a boom of families that can afford to stay in the city and pay for childcare rather than make the traditional flight to the cheaper suburbs.


  4. Pete Saunders

    Further evidence of the trends you cite — CTA year-over-year ridership is ticking upward despite the flat or declining population in the city. My guess is that trend is possible because of the dual income/dual degree households you mention. Overall population is down, but with greater numbers of households with two earners, instead of one, means higher ridership.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Basic HTML is allowed. Your email address will not be published.

Subscribe to this comment feed via RSS

%d bloggers like this: