Appalachia is Dying. Pikeville is Not.

Deep in Kentucky’s Appalachian Mountains, a Small City Finds Success

Regular readers of this blog know that I am always skeptical of “city success stories,” and most recently played a bit of debbie-downer on the accomplishments of Roanoke, VA. It will, then, come as a surprise to many that I really have unapologetic optimism about the city in today’s post: Pikeville, Kentucky.

Pikeville’s recent growth has been surprising in many regards, but is a testimony to a few key economic development principles. And when we consider the context in which Pikeville’s story occurs, the results are even more impressive.

So let’s step back a bit and look at Pikeville.

All pictures are from Google Streetview. Huge thanks to the diligent folks who’ve been photographing Pikeville — many of the pictures I use are dated as being taken in October 2016. Sadly, the vast majority of the roads around Pikeville have not been Streetviewed, and even many that have are now nearly a decade out-of-date. Several places do have recent pictures by the portable camera-orb or whatever it is, but not as many as could be hoped for a city wanting to promote its tourism.

But Lyman — I’ve Never Heard of Pikeville

Okay, fair enough. Pikeville is a small city of under 10,000 people. It’s not even part of a metropolitan area. Wedged into a narrow river valley in mountainous Pike County, Pikeville could hardly be further away from the great urban centers that urbanists love to write and read about:

It’s worth reiterating again that those gray splotchy bits you see in Appalachia are not, as you might expect, cities. They are coal mines.

When I say Pikeville is deep in Appalachia, I mean deep. You couldn’t get further into Appalachia if you tried. The terrain is rugged, steep, and beautiful. Farming is difficult, especially on the scale needed for modern mechanization. Mining has been the dominant industry for over a century.

We can also look at a map that gives us some information about transportation access:

The major interstates could not more completely avoid Pikeville if they tried. Even the traditional migration routes avoided Pikeville: down and left from Pikeville you’ll spot Middlesboro; that’s the Cumberland Gap, where migrants first pushed through the Appalachians into Kentucky. The mountains around Pikeville were too rugged for ready passage through.

The state highways to Pikeville are, of course, quite robust, often reaching interstate-quality. This is no accident. The Appalachian Highway System connects Pikeville, and one of the largest earth-moving projects in the western hemisphere occurred to reduce flooding in Pikeville and make way for a road. The Pikeville Cut-Through took over a decade, and moved 18 million cubic yards of earth. That is equivalent in absolute size to Boston’s “Big Dig.”

Here’s a fun view of Pikeville showing the size of the Cut-Through, on the left:

Thanks, Google Maps 3d Modellers!

You can also see there that Pikeville, like most geographically-constrained cities, is very dense. The Census Tract containing downtown Pikeville (essentially the whole middle of the above image), for example, has a density of about 2,687 people per square mile. While not incredibly high, it should be noted that fully half the land area in this census-tract is taken up by nearly-undevelopable mountains and hills. A more correct density is probably something like 4,000–5,000 people per square mile. This 4,000–5,000 person per square mile density, while not exceptional for an urban center, is exceptional for a few-thousand-person town in the middle of nowhere with no nearby cities to speak of. I looked at towns near where I grew up in Jessamine County for comparison. Downtown Nicholasville, which has no geographic impediments whatsoever and for which the entire tract is developed, has a density of about 3,700 people per square mile: but Nicholasville is more than twice as large as Pikeville is in total population. My hometown of Wilmore, just a bit smaller than Pikeville, has a density of about 2,200 people per square mile, which is comparable to the geography-unadjusted level for Pikeville.

In other words, certainly when you adjust for undevelopable area, and likely by any standard, Pikeville is denser than you’d expect for a city its size. The reason is blindingly obvious: there are mountains everywhere else, forcing more compact development.

Is this what you expect downtown for a city of 7,000 people to look like?

It’s not what I expect!

So, now that we have a sense of Pikeville’s geography, let’s turn to the question of population.

Looks Can Be Deceiving

Pikeville’s Population Record Actually Very Impressive

The above chart shows several population figures. The black line is based on official estimates of population for the Pikeville municipality. However, that black line needs to be taken with a grain of salt. Census Population Estimates for local areas can be very erratic. As you’ll see later in this post, this is doubly the case in small areas with many semi-rural populations.

Additionally, Pikeville has, at various times in its history, annexed surrounding territory. The result is that its definition as a municipality has shifted significantly across the time series, and so it’s not clear that the black line necessarily reflects a consistent area.

That’s why I give you the colored lines. Using decennial Census data, 5-year ACS samples, and a bit of hand-waving and extrapolation, I’ve developed annual estimates for tract-level populations from 1990 to 2013 for the Pikeville area. For the curious, “Inner Pikeville” is tract 9303/9903. “Pikeville surroundings” are tracts 93/9902/04/05.

I know what you’re thinking: Lyman, those lines don’t look very impressive. Sure, there’s growth, but not really that much.

But before I show you the really impressive graph, a few things to note:

  1. Inner Pikeville has grown in recent years. How many downtown mainstreets do you know of where the population in the absolute core census tract is growing again?
  2. Growth in greater Pikeville has been substantial in recent years, by anybody’s metric.
  3. Pikeville has had several periods of decline at various junctures, but has managed to turn it around and grow again after each period of loss.

These three factors alone make Pikeville’s population history interesting and at least a little bit impressive. But now for the big kahuna!

Pikeville is a growing city in a shrinking county.

I’ve written before about the long-run population decline of Appalachia. Pike County, one of the biggest coal-producing counties in the nation in many years in the past, should be ground-zero for that decline. And, in many ways, it is: outside of Pikeville, Pike County is at its lowest population level since the 1930s.

But, curiously, that weakness has not crippled Pikeville.

Let’s revisit the population growth of Pikeville, but this time we’ll compare Greater Pikeville to the remainder of Pike County, indexed for 1990. Due to data constraints I only have population estimates for these areas up to 2013 but, if the data from Pike County generally and Pikeville City are any indication, the trend below probably continued in recent years:

Here’s the great thing: because I’m using Census tracts that have had stable definitions over the whole period, we can be sure none of this change is due to Pikeville annexing territory from the county or anything like that. “Greater Pikeville” has managed to score 13% population growth since 1990 despite 19% population decline in the surrounding county. That is really impressive.

With a better understanding now of what we’re seeing, it may be worthwhile to zoom in on downtown Pikeville again, Census Tract 9303.

Looks a bit more impressive now, doesn’t it? Zoomed in, we can see that Pikeville has made a rapid recovery of population it lost from 1990–2000, with a population resurgence beginning around 2008 or so. The line ends at 2013, but Pikeville has almost certainly continued to grow.

You may be wondering why this population trend is the way it is. Let’s look into that question.

Pikeville’s Changing Economy

We can start where I usually like to start: with simple BEA data on personal incomes. From 1969–2014, I have data on personal income per capita for Pike County versus Kentucky and the US on the whole, as shown below:

So first off, we can see some coal booms in the 70s, 80s, and late 2000s. However, it bears noting much of the late-2000s increase is just that Pike County didn’t have as far to fall when recession hit as much of the rest of Kentucky or the nation on the whole.

We can also note that incomes have fallen relative to the rest of the nation since the recession. Pike County has been left behind. But even beyond that, it turns out that in 2014, nominal personal income per capita was unchanged from 2011. Nominal income has flatlined in Pike County, even as it was risen steadily in Kentucky and the nation on the whole. This is, obviously, because of declining coal production. Let’s look at that coal production.

Here’s total labor hours in eastern Kentucky coal mining:

Woah! Look at that dropoff! That’s incredible! It’s also incredibly sudden. Now there are a lot of reasons for this. Regulations on the coal industry, as well as carbon emissions regulations more generally, have hurt the coal sector, as have subsidies for clean energy. Meanwhile, cheap natural gas has dealt other fossil fuels, including coal, a body blow.

But the suddenness may also (and I say may because I haven’t seen definitive research on this, so consider it speculative) be due to unionism. Mines often were unable or unwilling to restructure their workforces easily, and thus kept large labor pools even as productivity per labor hour declined. What has happened in recent years is that many mines have entered into bankruptcy or other distressed conditions that enable them to aggressively renegotiate labor relations or terminate workers who were otherwise hard to fire. Had these mines been faced with a more flexible labor force a decade ago, it is possible that they might have shown a smoother transition in terms of demand for labor.

There’s another reason, however, that mining employment in Eastern Kentucky is declining:

Let’s talk about what this mine is and is not showing. What it is showing is that existing mines are getting close to being exhausted. What it is not showing is that there’s no coal left. Eastern Kentucky still has probably in excess of 45 billion tonnes of coal remaining.

There’s also a bit of an optical illusion here as well. A mine shutting down will reduce the “recoverable reserves at producing mines.” So if these mines went back to work, “recoverable reserves” would likewise tick upwards, as would productive capacity. Some mines could restart production fairly easily if demand picked up, which means recoverable reserves would jump back up. However, some mines shut down because their recoverable reserves were simply less and less competitive.

But nonetheless, while these lines exaggerate the rate at which coal reserves are being depleted, it is true that much of the best, cheapest, and most accessible coal has been mined out. Remaining coal mines have exhausted many of their best deposits, and getting at the reserves outside of existing mines requires the creation of entirely new mines, which is much costlier than continuing to work existing mines. Ergo, to open new mines, you need higher coal and energy prices, and the prospect of higher energy prices for a long time to come.

The depletion of many of Eastern Kentucky’s best, most accessible, and cheapest coal reserves does not explain the suddenness of decline in demand for labor at the mines, but it does suggest that any revival in coal production is likely to face great difficulty in sustaining any employment gains. Eastern Kentucky’s competitive advantage in coal mining is wearing out.

Up to now, then, I’ve given you no information about why Pikeville is doing so well. It’s been all bad news. And this is where it gets tricky: most mines are not in Pikeville, or even Greater Pikeville. A quick Google Earth perusal of Greater Pikeville indicates probably no more than 2–3 substantial mines anywhere within its territory. Pike County, on the other hand, has numerous mining operations, many of which have been impacted by the recent tumult in the coal sector, including Alpha Natural Resources, which declared bankruptcy (and came out of bankruptcy earlier this year). Thus, while coal is important in the area, the Pikeville city itself will feel these effects less than the county on the whole. Pike County is a very, very big county with quite a lot of people and many towns.

In Pikeville, we can find a different source of economic strength: education.

The University of Pikeville has been expanding rapidly. In 1997, it added a medical school. It has since added graduate programs in business and education, expanded its undergrad offerings, and is opening an optometry school within the next year. Now approaching a tripling of its size in just a few years, the University of Pikeville has drawn in literally thousands of new students, as well as faculty. And that’s not to mention the blue-collar workers who actually keep a university or hospital running day-in, day-out, or the increase in demand for construction driven by large university building projects.

Many readers will be surprised to hear me say that a small, Appalachian mining town is growing as a result of its robust educational sector. Popular stereotypes of Appalachians are of an uneducated people with little interest in higher education. The reality is that, to the extent Appalachia lacks higher education, it is mostly because of (1) poverty, or (2) geographic bias of university placement away from rural, mountainous, extractive-industry-intensive areas.

So far, UPIKE’s offerings are mostly focused on providing for the service needs of its community: primary care doctors, teachers, local business training, etc. This kind of talent-development and workforce training is useful and beneficial for the local community. But even more important than workforce development is knowledge-generation, and especially research, innovation, patenting, and publication. As the school has expanded, it has also begun to undertake more of this research and knowledge-creation.

Oh and, by the way: what happens when you add 2,000 students, who aren’t earning money, while only adding a limited number of jobs? Oh, right, personal income per capita goes down. At least some of the decline in personal income per capita is, then, a good thing, as the student population grows. Although students don’t earn much, they do consume quite a bit, and their tuition employs university workers, from white-collar administrators, to professors, to the blue-collar workers who keep buildings running, work in cafeterias, and secure university grounds.

It is not too much to say that the University of Pikeville is saving the city. Those 2,000 new students amount to essentially 100% of the growth in Greater Pikeville. This knowledge-and-talent-production facility in turn helps make Pikeville Medical Center a booming and effective employer, with a brand new building completed in 2014. Pikeville Medical Center is the only Mayo Clinic Network partner in Appalachia which, again, helps it operate as not only a hub for medical services but also a key knowledge-generation site.

UPDATE: A friend has raised the point that UPIKE’s rise is roughly similar to the sudden rise in student loan debt around the nation. His point was, essentially, that Pikeville may be riding just another boom that will soon bust, leaving the city disappointed. He also raised the question of whether Pikeville was the best place to invest higher ed resources. For the first question, I’d say that I do not personally think any coming higher ed “bust” will be anything close to as cataclysmic as the coal bust has been. But more to the point, UPIKE can do things to limit its exposure: keep net tuition at low levels, ideally rolling back some of the hikes in recent years. Start shifting more program focuses away from workforce training and towards research and knowledge creation. Work on expanding graduate programs even more, while keeping undergrad or part-time student growth more restrained. All of these strategies can help the school be better-prepared to weather any coming decline in student credit availability. For the second question, I would simply say that I think Pikeville is as good a place as any for higher ed work. Urbanists often favor urban locations for higher ed due to “agglomeration effects,” whereby proximity to industries, other universities, civic institutions, etc make research and workforce development more efficient for given inputs. To the extent agglomeration effects matter, Federal investments in UPIKE don’t make sense. On the other hand, knowledge generation is just one of the many positive externalities associated with universities. Strengthened civic institutions, a better educated local population, economies of scale from a larger population and consumption base, and the enormous benefits to neighbors of increased social capital are all likely positive externalities, and almost certainly have a bigger impact per input in a small city than a large one. So on the “optimal location of universities” question, I’d say we have to confess a degree of agnosticism about whether big, medium, or small cities are best for university investments on average. On the margin, however, small cities like Pikeville, especially Appalachian cities, have critical underinvestment in nearby higher ed services, so it seems likely that the marginal investment is more likely to be a good one here.

I got in contact with the Pikeville City Manager as well, who took the time to give me a lengthy response to some questions. He gave me a list of over 15 new restaurants to have opened in the last 10 years, with 5 more soon to open, as well as new hotels, retails, etc. Plus, he told me about their various revitalization efforts for the downtown, cultural attractions, a new distillery opening soon, etc. All of this is to say that, as Pikeville’s population grows, it is becoming more amenity-dense, which is a necessary precondition for lasting growth.

Much of this construction and investment has been helped along by Federal grants. Pikeville has been competitive at getting Federal, and especially Appalachian Regional Commission-related, money, as well as community reinvestment funds. Nobody will know for sure whether this money is well-invested unless and until the spigot is cut off, so I don’t want to sound too boosterish. Between hospitals, an expanded theater, a huge new conference center, a new city hall, drag racing, or the first Bourbon Trail stop east of Lexington (!!!! this one is a huge deal, readers. If you’re not from Kentucky, get with it and go on the Bourbon Trail, which will soon include Pikeville), some of these investments will turn out to be unsustainable boondoggles. But many will turn out to be very savvy, especially those that actually give Pikeville something to sell more widely: unique touristic experiences (which, for the record, having never been to it before, The Breaks is actually on my top-10 list of US parks I want to visit), knowledge generated by university research, medical services, craft and heritage products, etc.

And, at the end of the day, these investments would all be unsustainable if there was no hope for the population base. But there IS hope for Pikeville’s population growth.

And with that, let’s turn to demographics.

Appalachia Is Dying. Pikeville Is Not.

The Appalachian region has very high death rates and comparatively low birth rates. This is both because the region on the whole is very old, and also because it tends to be unhealthy. Middle-aged Appalachian men in particular have incredibly high death rates, which have risen especially as substance abuse problems have gotten worse. Pike County is no exception to this trend. And as marriage rates are not exceptionally high, fertility even for young women is not particularly high.

But within Pikeville, the story is a bit different. The data available is not extremely precise, but it appears that Greater Pikeville’s birth rate is substantially higher than the rest of Pike County — and its death rate is somewhat lower. This is both because Pikeville’s population of young women is expanding as the University of Pikeville recruits students, and also because it seems people in the city are just more fertile and slightly less likely to die.

If anyone is curious about the exact nerdy math on that, I can share it, but the key takeaway for the non nerds is this: Pikeville can grow by natural population increase. Pike County cannot. This means that for Pike County to even maintain its current population, it must attract net positive migration, something it only ever does during big coal booms, and even then not by very much. But all Greater Pikeville has to do to grow is to break even on net migration. The city doesn’t even need positive net migration to grow.

To demonstrate this, I have simulated a range of birth and death rates for Greater Pikeville that roughly approximate the error-ranges I observe in ACS data, and that follow some recent trends. This gives us a good “error band” for the population of Greater Pikeville. I also include a forecast made by, I believe, Kentucky’s State Data Center in 2005 for Pikeville’s population; these local area forecasts have apparently been discontinued (because they are super-duper hard to do). So, here’s a range of forecasts for Pikeville’s population out to 2030:

As you can see, the 2005 projections were way off base, but that’s partly because Pikeville annexed a lot of territory into itself, so we shouldn’t be too hard on those forecasters. The goal-posts moved, so it’s not their fault.

When we look at the standardized Greater Pikeville area, we can see that there’s a wide range of possible outcomes. In every scenario, I have assumed that Greater Pikeville’s net migration is 0, i.e. it gains just as many people as it loses, breaking even. I have also assumed that Pikeville’s age-and-sex demographic composition changes at a normal rate; if you wish to assume demographics will shift more slowly, then just narrow the endpoints a bit.

The point is: the central tendency is for Pikeville to continue to exhibit growth. My estimate is that population would be about 16,700 in 2030, up from 16,300 in 2013. Now, that’s pretty slow growth. But that right there tells us something: we’ve got better-than-even-odds that Pikeville experiences population growth, simply based on recent demographic trends.

If you add in a few friendly assumptions like (1) UPIKE keeps expanding, (2) Pikeville attracts a few more substantial employers (their city manager claims they already have several big employers fixing to move in soon), (3) improving job, amenity, and educational opportunities reduce outmigration, or even (4) coal production makes a short rebound, then my higher forecast of about 18,400 people in the Greater Pikeville area starts to look attainable. It would only take a very moderately positive net migration rate to get you there.

ACS data are such that I cannot get net migration rates for census tracts. But I can look at inflows. This allows me to see how inflows vary in Greater Pikeville versus the rest of Pike County. The results are telling:

Even as fewer and fewer people are moving into Pike County, more and more people are moving into Greater Pikeville, at least as of 2013. This would seem to suggest that we have cause for optimism for Pikeville’s demographic future. The UPIKE enrollment data goes up 3 years past the Greater Pikeville population data, and suggest that population will keep rising. Other improved economic and social conditions may yield further follow-on-effects.

It seems highly likely that Greater Pikeville will include at least 18,000 people by 2030.

How many of this people will reside in Pikeville proper is, of course, unpredictable. It will depend on local zoning codes, commuting patterns, where amenities locate, and patterns of city annexations. I won’t venture a prediction of so unstable a target. But the Greater Pikeville figure indicates that the region on the whole stands to reap real benefits.

Conclusion

Pikeville is defying the odds and experiencing real urban growth in the midst of Appalachian population collapse. This is a huge success story with lessons for many communities. The success of small cities is based on their ability to draw people in with jobs, amenities, and low costs: universities serve as excellent incubators for at least the first two. Moreover, universities need not be public, nor enormous. UPIKE is a private school and isn’t huge. But university and government leadership have both invested in their communities, pouring energy and resources into assets: knowledge creation, civic organizations, institutions, reputational capital, etc. Furthermore, Pikeville’s success shows that post-industrial towns deep in the coal belt totally isolated from the large urban centers can nonetheless succeed. Their success will look different. They may never take home the fat paychecks of Silicon Valley. But they can nonetheless have functional, economically viable towns that give their young people a shot at achieving their dreams, often in their very own hometowns.

It remains to be seen whether Pikeville’s success will endure: new academic programs sometimes fail, government grants can create huge operating costs for local governments, and returns from business recruitment and tourism can prove fickle. We should bracket optimism for Pikeville’s future with realism: every new asset has operating costs, and those operating costs require revenues. Pikeville’s city taxes are high relative to most of its neighbors and Kentucky on the whole and, like most Kentucky government units, licensing, bureaucracy, and regulation continue to be burdensome, both weakening local businesses and providing avenues for government corruption and cronyism. Pikeville’s success suggests these problems are being well-managed thus far; but it will take continued skilled management to keep the city’s successes going. I did not analyze Pikeville’s municipal finances; I do not know if the city has the financial capacity to manage its financial burdens in the long run. If I get positive feedback from this post suggesting there’s an interest, I will take a look at that question.

But for now, it’s enough to say that Pikeville is growing. Ask almost any urbanist in the country if a small town in Appalachia with no interstate, no major airport, a history of coal dependence, and a small Presbyterian college is likely to grow and they’ll all tell you the same thing: heck no! Pikeville is beating the odds. Here’s hoping the city can keep it up.

There are things Pikeville can do to improve its current good trajectory. One small thing I noticed is that the whole city has a whopping total of one AirBnB listing, and it’s pretty far from Downtown. AirBnB is a great way for locals to directly benefit from any increase in tourism, university event attendance, etc. Likewise, Pikeville could consider reaching out to AirBnB about how to do a pilot run on AirBnB’s new “City Experiences.” AirBnB has focused on large cities, but Pikeville might consider seeing if AirBnB is interested in experimenting with including listings for small cities. These aren’t huge changes to propose, but they’re just things to think about for Pikeville’s leaders.

And, as always, if you’re reading this thinking, “Gee, I wish Lyman would take a look at my city,” you need only fire up the Twitter-machine and Tweet at me. My only fee for my services is that you agree to use any information I provide to make your town a better place.

You can see a community reinvestment project in Pikeville here.

A fun, if a few years out of date, video showing the community here.

Check out my Podcast about the history of American migration.

If you like this post and want to see more research like it, I’d love for you to share it on Twitter or Facebook. Or, just as valuable for me, you can click the recommend button at the bottom of the page. Thanks!

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!

I’m a native of Wilmore, Kentucky, a graduate 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.

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. More’s the pity.