Kentucky’s Migration History
Comparing the American Community Survey and the Population Estimates Series Shows Differences
I’ve written extensively in the past about migration in my home commonwealth of Kentucky, including deep-dives into the regional migration profiles of Inner Appalachia, the Western Pennyroyal, Greater Louisville, Greater Cincinnati, and the Bluegrass. For those pieces, I primarily used the American Community Survey’s 5-year sample from 2008–2012. The ACS is a great source for doing detailed analysis of specific regions, but it has one big shortcoming: it doesn’t go back very many years. For the level of detail I want, it really doesn’t go back before 2008. However, the Census Bureau produces standardized estimates of population change for every county in the United States, called their Population Estimates series.
The PE series is the Census Bureau’s “best guess” at the population and basic demographics for each county in the US. They announce their preliminary numbers for each year around April or May of the following year, so, as demographic data go, they are very, very quick. The data is easy to use, well-formatted, and has the Census Bureau’s official seal of approval, so it’s understandably popular.
For today’s post, I’ll use the PE series to look at Kentucky’s migration history… and also use Kentucky’s migration history as a way to show the strengths, and weaknesses, of the PE series. As we will see, regional migration profiles can look quite different depending on what data series we use.
Kentucky’s Population History
Statewide Population is a Boring Story
With the exception of a downward blip for WWII, Kentucky’s population growth has been fairly linear. Sure, there have been periods of faster or slower growth, but generally this is not an extremely interesting story. But what if this statewide trend is hiding a good story?
Now, full disclosure: for statewide population, Census produces official intercensal estimates of total population going back to 1900. I use those numbers. For the county-level data I’m going to show below, I only have those intercensal figures going back to 1990, and have to use imputed data based on decennial censuses for the interim. I adjust imputed intercensal regional populations from 1900–1990 based on the annual statewide population, but from 1790–1900 intercensal population volatility is unavoidaly eliminated.
Kentucky’s Population History
Regional Population is Volatile
Using regional population estimates, several stories appear. First, Inner Appalachian population growth was explosive from around 1900 to 1940, on the back of the coal boom, electrification, and some industrialization of the countryside. The region went from the least-populated part of the state to vying for the second-most-populous position. Since 1950, however, Inner Appalachia has seen population decline, and may be the least populated part of Kentucky again in another decade or two.
Another story that stands out relates to WWII, mobilization, and urbanization. Different regions recovered from WWII differently. Now, we have to treat this data carefully, because the annual changes are based on average growth rate between Censuses (1940,1950, and 1960 in this case), adjusted for the statewide growth rate. But presuming that a vector of statewide growth and regional intercensal growth is a good way to estimate annual population, we can see that WWII mobilization had differential impacts across the state, even when I distribute the direct recruitment effect evenly.
After WWII, urban areas like Lexington, Louisville, and Cincinnati exhibit a radically different population trend, sharply upwards. At the same time, most rural populations did not recover from WWII for decades. Once the GIs came home, they tended to gravitate towards urban areas, and the mobility and educational uplift (and recovering economy) facilitated by the numerous transitions after WWII kicked off the modern wave of urbanization (and suburbanization). But the key point is this: many researchers locate the source of sudden urban and suburban growth in post-war decisions by policymakers and individuals. But this data suggests that the war itself was an essential population reshuffler: it yanked millions of young men (and, directly or indirectly, young women as well) out of rural communities and flung them around the world, raising their future propensity to migrate and seek residence outside of their rural communities.
Kentucky’s Population History
Recent Population Trends
The long view has some interesting stories, but the really interesting focus is on more recent years. Since 1990, regional population growth rates have diverged widely. The result is that Inner Appalachia’s population has shrunk by about 13 percent, while the KY-side of the Cincinnati suburbs have grown 35 percent, and the Bluegrass area 36 percent. These are jarring changes. Cincinnati and Lexington have seen huge construction booms, suburbanization, and economic vibrancy, while the mountain counties of Inner Appalachia have seen emptying towns, shrinking churches and schools, and a growing sense that things are getting worse.
Kentucky’s Migration History
Domestic Migration Rates Vary Widely
Domestic migration rates vary around Kentucky. For Inner Appalachia, they are deep in the red. For the Bluegrass or Louisville, domestic migration is a more positive force. But across Kentucky, domestic migration has declined in recent years. From attracting a net of 15,500 people in 2005, in 2012 Kentucky lost 5,500. This decline was most extreme in Appalachian counties and Northeastern Kentucky, but every region saw significant declines.
Now, in my previous writing on Kentucky, I’ve come up with maps and charts to show where migrants are coming from and going to. But I can’t do that with the Population Estimates series, because all I have is net migration. Plus, even if the PE series gave gross flows (so that we could see which regions have high outflows versus low inflows as a source of declining migration), I still couldn’t see where migrants go, because the PE series is a model result, not strictly empirical data. The PE series uses data from administrative records like the IRS and the military, survey data from the American Community Survey, and many other sources to produce an estimate, but that estimate has no direct tie to any specific piece of data. That being the case, though the PE series shows declining migration, it can’t tell us much of anything about who left, who arrived, how that changed, and why it changed.
Kentucky’s Migration History
Domestic Has a Large Impact on Population
The above chart shows an estimate of the “cumulative effect” of annual net migration. It isn’t really a strictly meaningful demographic estimate, but rather can serve as a broad indicator of which regions have, over the last 25 years, seen the most and least net migration. And the story is clear: Inner Appalachia has seen net out-migration with a direct impact equal to about 13 percent of the population. Meanwhile, the Bluegrass, Greater Cincinnati, and the Eastern Pennyroyal have seen net in-migration equal to about 12 percent of the population. These are just direct, unadjusted impacts, and should not be taken too literally(in reality population changes in the long run are as much impacted by the composition of gross migration as the amount of net migration), but can be indicative.
Kentucky’s Migration History
International Migration is Positive (Almost) Everywhere
International migration tends to be fairly stable… with a few exceptions. First of all, Census has no single, reliable source for data about emigration. So instead, they impute emigration using a variety of methods. However, for emigrant groups about whom they do have good information, they input it fairly directly. So, for example, the Western Pennyroyal has two large military bases, both of which deployed soldiers abroad in 2002 and 2003. The result: negative international migration.
Furthermore, Census has (understandably) struggled to figure out how to measure migration accurately. So you’ll note that migration rates rise everywhere from 1999 to 2001? I don’t know for certain, but I’d say there’s a strong chance that’s due to a change in methodology, not necessarily an actual rise. Again, the PE series isn’t one dataset that has been collected systematically for decades: it is a model result that arises from shifting underlying datasets.
International migration is highest in urban areas, once military migration is excluded. Jobs, cultural or ethnic ties, education, and easier access to information make urban areas far more likely to draw international migrants than rural areas.
Kentucky’s Migration History
International Migration is Positive (Almost) Everywhere
As can be clearly seen, two regions stand out above the rest. Lexington and Louisville have been the largest beneficiaries of international migration in the Commonwealth for over 20 years.
Measuring Migration
Why Estimates Disagree
Regular readers will recall that my previous writing on Kentucky has suggested that net migration out of Appalachia is probably overstated. That is because I looked at ACS data, and specifically 2008–2012 ACS data. Above, you can see that 2008–2011 at least was a very good time for Inner Appalachian migration, almost becoming positive, although it has become more negative since. So on some level, my more bearish outlook on Appalachian migration here reflects different time horizons.
But there’s something else at work too.
To put it plainly, ACS and PE estimates are not even in the same universe as one another. They don’t even agree on the net direction of migration, even for big regions like Louisville. It’s almost comical how wildly different the estimates are when you zoom in to small regions.
So, if their estimates are so different for the regions… might the whole statewide net migration estimate be wrong?
There’s two ways to look at the graph above. On the one hand, we could say, “both series show a decline in migration,” in which case we would tend to say they agree. But on the other hand, we could say, “the year-over-year direction of change and amount of change are frequently different between the two series,” in which case we would tend to say they disagree. It’s hard to say which reading is necessarily better.
But you’ll note I included a third data series: the IRS Statistics of Income data for migration. This number matches the Population Estimates series much more closely. Fancy that.
The reason, of course, is that the PE series uses IRS data as an input data source. ACS data is not used in the same way, despite numerous limitations and shortcomings to IRS data. Because IRS data has been around longer (and ACS data just in the last few years could have been a viable substitute), it is used much more consistently. Plus, as “administrative” rather than survey data, it’s generally considered to be more reliable and available (personally, I trust surveys more than administrative data).
So ACS data gives us different estimates of migration at every level of geography. That implies different patterns of migration flows. Depending on what data source you trust (and depending on how you define migration: ACS generelly captures more different forms of human mobility, migratory or otherwise, such as daily commuting patterns), you can get very different estimates of regional migration patterns.
By any measure, Kentucky’s net migration rate has worsened over the last 5 to 10 years. The ACS still shows positive migration, while the PE series shows net outflows, but both show declines. Which regions have the strongest or weakest performance largely depends on which data source you trust. Where the ACS and PE series agree is that the Bluegrass leads the way on domestic migration, while Outer Appalachia lags behind. Over the course of Kentucky’s history, successive waves of migration have altered the commonwealth’s total population, and that population’s regional distribution. From early settlement to plantations to industrialization to the coal boom to modern suburbanization, migration trends have made Kentucky what it is today. That reality will continue, though it’s far from clear what tomorrow’s migration will look like.
See my previous posts, about nationwide migration history and recent trends.
See my previous posts about Kentucky.
Start the series from the beginning!
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 graduate of the George Washington University’s Elliott School with an MA in International Trade and Investment Policy, and an economist at USDA’s Foreign Agricultural Service. I like to learn about migration, the cotton industry, airplanes, trade policy, space, Africa, and faith.
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.
