The best part about Cincinnati is your can see Kentucky from it. Photo by Robert Conklin on Unsplash

The Rust-Belt Comeback, Cincinnati-Style

Not Pittsburgh.

Everybody is used to talking about how Pittsburgh has halted its Rust Belt decline. But I’m actually not wildly impressed, and there’s a different Rust Belt city that I think really may be a good statistical candidate for turnaround: Cincinnati.

Now, full disclosure: I’m biased in favor of Cincinnati. I have family ties there, and have family who work for an organization whose fate is bound up in the wellbeing and national attractiveness of the city. I’m an enormous fan of Cincinnati’s cultural heritage, from goetta to Graeter’s. Whenever my wife and I visit, we buy 5–10 pounds of goetta to bring back to DC. So, to be clear, I’m super biased.

But let’s take a look at the actual data. Let’s start with the entire metro areas, with current metro-area designations back-cast to 1800.


The two areas had similar populations through the 1880s. They were both big hubs for industrial activity, both received heavy immigration, both heavy on transportation and heavy industries. Both enjoyed the benefits of being close to coal in particular. But Pittsburgh was closer, enjoyed even larger-scale industrialization, and Cincinnati lagged somewhat.

But with the Great Depression, things changed. The Pittsburgh area seems to have been hit harder than the Cincinnati area. And then wartime mobilization hammered Pittsburgh hard, while Cincinnati barely gets a blip. Post-war, Cincinnati’s growth is incredible, with the region adding half a million people, or about 40% growth, in the span of a decade. Cincinnati’s fastest growth was in the suburbs: Boone County, Kentucky, Clermont and Warren Counties, Ohio, but solidly half of all population growth in the period was in Hamilton County proper.

By the late 1960s, growth had slowed in Cincinnati, and gone negative in Pittsburgh. To this day, Pittsburgh has not seen any sustained recovery in population, while Cincinnati’s growth has continued steadily, again, especially in its suburbs, and over the river in Kentucky. Within a decade or two, Cincinnati will probably be a bigger metro area than Pittsburgh.

But suburbanization isn’t what people mean when they talk about the urban inversion, or urban renaissance, of our current period. We want to know if more local geographies nearer the city core are prospering.

So let’s look at just the primary counties in question.


Again, there’s rough parity through the 1880s. Then Allegheny County leaps ahead. Both have a slower Great Depression, but Allegheny County is hit worse. Hamilton County has a stronger war and post-war experience… but then decline sets in about the same time it did in Allegheny County.

The trick is, however, that the decline is far less severe. By my count, Hamilton County in 2017 was probably 12% below peak population, while Allegheny County was 25% below peak population.

This is shocking! Allegheny County has about 730 square miles of land, vs. Hamilton County’s 400. Pittsburgh is in the middle of Allegheny County, whereas Cincinnati is on the edge of Hamilton County. Suburbanization should have hit Hamilton County far harder than Allegheny County! Escaping Hamilton County to move out to a new jurisdiction is a lot easier than escaping Allegheny County. That is to say, the average person suburbanizing out of Allegheny County probably had to move a lot further and extend their commute by a lot more than the average person suburbanizing out of Hamilton County.

And crucially, it seems like Hamilton County is now growing again, while Allegheny County is just stable. I say seems because the post-2010 data is still in flux. I use vintage 2016 population estimates as a base, but then take the county-level share of vintage 2016 state population, and multiply it by vintage 2017 state population. For the year 2017 itself, I assume the trend in county share of population continues.

As a result, my data suggest that, from a low of 799,000 people in 2008, Hamilton County’s population has probably risen to 810,000 or more people in 2017, and the growth appears to be chugging along steadily. Allegheny County, meanwhile, grew from 1.221 million people in 2008 to about 1.233 in 2013… but I believe it has since slumped back under 1.23 million people. Its boom seems less robust than Cincinnati’s.

We can check this other ways too using various high-frequency data. For example: housing construction permits in Allegheny County and Hamilton County.


Allegheny County was beating Hamilton County handily. Now it isn’t. One reason why? Here’s rents:


Rents in Pittsburgh are falling. Rents in Cincinnati are not. This, despite the fact that Cincinnati has an enormous amount of vacant real estate being actively renovated and flipped, and that the metro area includes a wide range of jurisdictions, many with very generous zoning rules. In other words, Cincinnati’s rising rents are probably not mostly due to supply restrictions.

We can also look at Google search volumes. Here’s searches for housing in the two cities:


There is insufficient data to do “move to X” for Cincinnati, so we can’t make that comparison, sadly.

We can also look at a sort of silly indicator but one that shows yet another case of Cincinnati’s stronger recent performance: arrivals at the major airport!


Most of what you’re seeing here is changes in the structure of the airline market, with CVG and PIT losing major airline hub status. But in recent years, we’ve seen CVG gradually reclaim its share and attract new airlines. A whole passel of discount airlines now operate out of CVG, as does Amazon’s air mail service. The high density of new airline routes opening out of CVG is partly due to Delta legacy overcapacity, but also suggests that airlines are observing demand there. Representing this as a share of national arrivals helps control for the business cycle, so this isn’t just a general national recovery in flying. Now PIT is also showing a recovery, but CVG has steadily narrowed the gap since 2014.

Which is to say, the wider Cincinnati area is probably a solid bet for best-performing Rust Belt-decliner. There are MSAs with better recent performance like Indianapolis and Columbus, but they either never experienced the large-scale industrial decline that cities like Detroit, Pittsburgh, or Cincinnati experienced, thus aren’t properly “Rust Belt,” or have zero-sum advantages like having a state capital. One of the other best-performing Rust Belt MSAs is Louisville, which has a lot of commonalities with Cincinnati.

But counties are still a big unit. Let’s zoom in to the municipality.

Cincinnati Proper is Growing Again

That’s Weird.

I looked over the records for Chicago, Detroit, Minneapolis, St. Louis, Cincinnati, Cleveland, Columbus, Indianapolis, Milwaukee, Louisville, Toledo, Pittsburgh, Dayton, Akron, and Youngstown. Of these, the only growing municipalities were Indianapolis, Columbus, Minneapolis, Louisville, and Cincinnati.

Here’s Cincinnati vs. Pittsburgh.


At current growth rates since 2010, Cincinnati could overtake Pittsburgh as early as the 2020 Census, though the mid-2020s are more likely. This would be the first time since the 19th century that Cincinnati has had more people than Pittsburgh.

The lines make it hard to see, but Cincinnati’s low ebb looks to have been in about 2011, at around 296,000 people. Since then, it has now recovered to about 300,000 people. Pittsburgh, meanwhile, hit its low ebb in 2016 at below 304,000 people, and only rose slightly in 2017.

Cincinnati’s growth is not, by any means, a massive boom. But it’s not nothing. Furthermore, much of this growth is in Cincinnati’s denser areas. The core downtown, business-district Census tracts saw population rise from an estimated 29,735 in the 2007–2011 ACS 5-year sample, to 31,228 in the 2012–2016 sample. Only 500 more people… but that’s with the very central business district losing 1500 people, I’m assuming due to some major apartment complex closing.

In particular, the Over-the-Rhine neighborhood (AKA Census Tract 16 for Hamilton County, Ohio) has seen its estimated population rise from 760 people to 1048 between these two ACS samples. ACS is probably underestimated OTR’s population in both cases, but the trend is probably right. Once a thriving German neighborhood, and then later a blighted district, OTR has been the focus of an intensive revitalization effort that, anecdotally and in the data, appears to be working.

If we include some of the OTR border-tracts out towards Washington Park, Sycamore Street, and Jackson Park (i.e. Census Tracts 9, 10, 16, and 17), we get an estimated change from 3,163 people to 5,208. That’s a big change! Big enough that the before/after populations are outside each others’ margin of error as identified by Census, which means we can be reasonably confident that there really has been a population increase in these neighborhoods.

Of course it could just be gentrification. But the key here is that while there have undoubtedly been inflows and displacements, the standard story of gentrification is of population parity or even decline. Usually, we expect wealthier inflows to want more space per person, and so inflows don’t outnumber displacees. That’s the case in many gentrification cases like in San Francisco or New York, or even DC: by and large, the gentrifiers tend to be of similar or less number to the displacees.

That’s not what’s happened in OTR. At least in this decade, the gentrifiers greatly outnumber the displacees. This is as you’d expect given the dilapidated condition of housing in some of these neighborhoods: comparatively few people actually reside there, and buildings must be overhauled to make them functional for a larger population.

Throughout Hamilton County at least, it turns out that tract-level density of the built environment in 2010 is a decent predictor of population growth.


This is the sort of thing we expect to see if a city is experiencing a broad-based improvement in the sorts of conditions that make people want to move there. Growth for the metro area, growth for the county, growth for the urban core. We expect to see blighted neighborhoods near the core get turned around. Oh, and we expect employment trends to start looking better. Here’s employment for the two cities:


There’s a concerning blip at the end for Cincinnati. But from 2010 to mid-2017, Cincinnati is crushing it, regaining essentially trend employment growth before a dip in late 2017. Below, you can see the difference between Hamilton County’s unemployment rate, and Allegheny County’s:

Hamilton County has had a higher unemployment rate than Allegheny County effectively forever. But the point is that that relationship has changed over time. Throughout the boom-years of the 2000s, the gap was 2–2.5%. But by 2016, the gap was just 1%. It’s now risen to about 1.5%, but, still, Cincinnati’s relative performance vs. Pittsburgh is, once again, improving.

Why Is Cincinnati Growing?

I don’t want to spend too much time speculating on this as it can be a dense topic and I really don’t intend this piece to be too voluminous. But I want to note something very straightforward.

The story about Pittsburgh is that high-achieving educational institutions, in partnership with Silicon Valley spinoffs and startups, are saving the city. I would suggest that if that’s true, they’re doing a bad job of it. Cincinnati, without those assets, seems to be making a faster turnaround.

But… is Cincinnati really lacking those assets? I’m not so sure.

Using IPEDS data, we can look at the 12-month unduplicated headcount of students attending 4-year public or not-for-profit universities in the Pittsburgh and Cincinnati MSAs. Here’s the graph:


The two MSAs actually have very similar numbers of students, despite Pittsburgh being a somewhat bigger MSA. Pittsburgh’s enrolled population is falling, while Cincinnati’s is stable. Pittsburgh’s decline has been heaviest in the California University of Pennsylvania, a public state college, Carlow and Duquesna Universities, two private schools, Geneva College (private), and the various Penn State University and University of Pittsburgh campuses. Other schools, like Seton Hill University, Slippery Rock University of Pennsylvania, and, of course, Carnegie Mellon University, are growing. But while Carnegie Mellon gets a lot of attention, it only accounts for less than 15% of students, and is not offsetting the broad decline in many public universities and less-prestigious private schools around Pittsburgh.

In Cincinnati, the story is a bit different. The flagship University of Cincinnati is growing steadily, while the suburban/exurban Miami University is also holding its own. There’s a variety of experiences among smaller private schools: some like Thomas More are growing, while others like Xavier are shrinking. And one of the largest nearby universities, Northern Kentucky University, is shrinking. But on the whole, the growers offset the shrinkers, so the metro area has a steady supply of talent.

Curiously, this trend isn’t confirmed by the American Community Survey. The 1-year ACS files show Cincinnati’s share of the population with a high school degree and enrolled in school to be pretty much identical to Pittsburgh. And even odder, both show a much higher enrollment rate than IPEDS shows. However, this may be due to online schools, or schools I do not include in the IPEDS samples. Any student taking online classes at a school not in the metro area, IPEDS data would miss. Any student taking classes at a 2-year school or a for-profit school, my IPEDS data would miss as well. These seem like plausible explanations for the difference: maybe Pittsburgh has more distance-learning, for-profit, or 2-year enrollments than Cincinnati does.

I’m inclined to take the IPEDS data. In the ACS data, I can’t even be sure I’m not catching remedial GED work or something like that. The IPEDS data, however, shows that Cincinnati’s “university core” is bigger and more stable than Pittsburgh’s.

Now, where there is a measurable difference is in graduate work. IPEDS data gives us the 12-month headcount of graduate students by MSA.


This is what people are talking about when they talk about Pittsburgh’s advantages: it’s got a big research sector. Bigger than Cincinnati’s by a substantial margin, which is a very recent development. The trouble is, though, that it sin’t actually leading to gangbusters growth, whereas’ Cincinnati’s more stable graduate share is occurring alongside actual population growth. In other words, if all you’ve got is graduate programs, it’s probably not enough.

So why is Cincinnati doing so much better?

Well, one reason is a better core demographic balance! The Cincinnati area averages about 27,000 births per year, which has risen since 2010, versus about 20,000 deaths. Pittsburgh averages about 23,000 to 24,000 births per year, and falling, versus about 27,0000 deaths. So Pittsburgh has to attract migrants just to have stable population, whereas Cincinnati can lose some people, and still grow.

This isn’t just a product of age. Using ACS data, we can get a general idea of total fertility rates by MSA.


As you can see, Cincinnati really does have higher fertility than Pittsburgh, even controlling for age. Now, this doesn’t mean women in Pittsburgh will never have as many kids, it just means that, for some reason or another, a woman of a given age in Cincinnati is more likely to have a kid than a woman of a given age in Pittsburgh.


This chart shows the cumulative distribution. As you can see, Cincinnati’s birth advantage starts in the teen years, around 16–18. That’s not a great sign actually, teen pregnancy is not something we want, as a general rule, to encourage. The gap keeps growing until the late 20s, though is mostly complete by age 25. By age 30, women in Cincinnati and Pittsburgh have about the same odds of having additional kids.

So most of this difference is about fertility between ages 18 and 25. College years. But it turns out, when I introduce a control for enrollment, the effects don’t vanish!


This chart shows the age-specific cumulative ASFR gap, Cincinnati minus Pittsburgh. Basically, it shows the accumulated difference in fertility for enrollment groups.

Cincinnati’s unenrolled women have much higher fertility than Pittsburgh’s unenrolled women, and the difference starts early. So Cincinnati’s 15–18 year old higher fertility can be largely written down to differences in fertility among women who are not in school.

But by age 19 and into the 20s, enrolled women in Cincinnati also have more kids. Here we are mostly talking about college or graduate students, and we see that they manage to “have it all” (i.e. school and kids) more than their Pittsburgh peers. Unenrolled womens’ fertility continues to rise as well.

By the mid-30s, Pittsburgh enrolled women have caught up, but unenrolled women haven’t. And to be clear, by age 35, the enrolled share of women is small: under 10% of women. In other words, the losses that 15–25% of Pittsburgh women experience vs. Cincinnati women from ages 20–25 are only made up by the 5–10% of Pittsburgh women who are still enrolled in their 30s, i.e. second-career people or women pursuing advanced degrees.

So it’s not schooling driving these differences. It’s just that Cincinnati women have more kids. One factor driving this could be marriage. Are Cincinnati women more likely to be married?



This is sort of an odd chart, but it’s showing something important. I take the married share of the female population for each year of age, averaged 2012–2016. I then add up the cumulative shares from age 15 to that age (say, 25). Then I divide the sum of those percentages by the number of years elapsed (in this example, 10). This yields the average share of her fertile years a woman who is 15 can expect to be married by the time she reaches the age shown, assuming age-specific marriage rates are unchanged. In both cities, women can expect to spend substantially less than half of their fertile years married. By the time a woman is 30, she can expect to have spent about 15–18 percent of her fertile years, or about 2.2–2.8 years, married.

Marriage is a strong predictor of fertility. So if women spend more time married, they will probably have more kids. Here’s the age-specific ratio in cumulative share of fertile years expected to be spent married, regressed against cumulative expected births.


So what explain’s Cincinnati’s higher fertility? It’s not that women do less schooling or aren’t getting what they want out of life. It’s that they spend a little bit more of their life as married women.

What about deaths?

Well, that’s trickier. I don’t know how to getting MSA-specific age-adjusted death rates. But we can do something simple. We can get the age composition of Pittsburgh and Cincinnati and apply national age-specific death rates, and use that to estimate what deaths should have been.

The chart below shows predicted and actual deaths:


In both cities, actual deaths substantially exceed predicted deaths. This could reflect a mismatch in ACS and CDC data. It could reflect that these happen to be two high-age-controlled-mortality places. It could be something else. But if we presume the cause of the raw gap is something stable and not germane to our analysis, then we can take the ratio between actual and predicted deaths as an indicator of the extent to which age-controlled death rates exceed national averages. While the exact ratio may be wrong, the trend should be about right.


In 2011, Cincinnati probably had a worse age-adjusted death rate than Pittsburgh, both relative to the country on the whole. But Pittsburgh has rapidly gotten worse, at a pace much faster than national age-controlled death rates, while Cincinnati has been nearer to something like the national trend. As of 2016, it looks like Pittsburgh’s age-controlled death rate is probably something a lot like Cincinnati’s, and it may get worse by 2017.

On the whole, then, we can see that a substantial part of Cincinnati’s strength relative to Pittsburgh is driven by natural increase. Thanks to higher marriage rates and a less-steep increase in age-controlled death rates, Cincinnati’s population is chugging ahead, while Pittsburgh languishes.

What About Migration?

Well, what about migration? Here’s my estimate of net migration for each metro area, 2000–2017:


My estimate is that, in 2017, Cincinnati may be getting more people from net migration than from natural increase. Much of this is in the wider suburban area, but a not-insignificant part of it is within Cincinnati proper. I personally know one young family that moved into northern Kentucky from New York City around 2014, but then moved into Cincinnati proper in 2017. Their social circles seem pretty full of that type of story.

The curious thing here is how different Pittsburgh and Cincinnati look. Pittsburgh throve during the recession. Cincinnati shows up, however, as losing people! This, despite the fact that, as I showed earlier, the denser parts of Hamilton County were gaining people. It’s just that Cincinnati has so much suburbia, its losses during the subprime crisis were quite large.


I’m bullish on Cincinnati, and it’s not just because I really, really love goetta. Cincinnati continues to have near-replacement-rate fertility, largely because the local community succeeds in getting people hitched. Age-adjusted death rates are not showing as severe of spikes as other midwestern locations. Net migration is looking pretty good too. There’s been a recent dip in employment, but hopefully that will turn around. Certainly there’s demand for people to live in Cincinnati. Major structural liabilities, like blighted neighborhoods, are rapidly being turned around. And it’s all happening fairly under the radar. Cincinnati isn’t getting the media attention of other turnaround stories, aside from people talking about the Over-the-Rhine neighborhood.

Other metros in the midwest like Indianapolis or Minneapolis are seeing faster growth, in no small part because they never saw as much decline. But in Cincinnati, we have our strongest case of a deep-decline city making a real turnaround. Here’s hoping it lasts!

I’m an an Advisor at Demographic Intelligence, the nation’s leading producer of rigorous national- and regional birth and marriage forecasts. I’m also a Research Fellow at the Institute for Family Studies, a Senior Contributor at The Federalist, and an I write periodically for Vox’s Big Idea column. I’m a native of Wilmore, Kentucky, a graduate of Transylvania University, and also the George Washington University’s Elliott School. My real job is as an economist at USDA’s Foreign Agricultural Service, where I analyze and forecast cotton market conditions. I’m married to a kickass Kentucky woman named Ruth. I am not paid one penny by anybody for this blog post.

Follow me on Twitter to keep up with what I’m writing and reading. Follow my Medium Collection at In a State of Migration if you want updates when I write new posts. And if you’re writing about migration too, feel free to submit a post to the collection!

Check out my Podcast about the history of American migration.

DISCLAIMER: My posts are not endorsed by and do not in any way represent the opinions of the United States government or any branch, department, agency, or division of it. My writing represents exclusively my own opinions. I did not receive any financial support or remuneration from any party for this research.

Like what you read? Give Lyman Stone a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.