The latest Census Bureau population estimates. Courtesy of the Census Bureau.

“Once more unto the breach, dear friends, once more.”

– Henry V, rallying his troops at the siege of Harfleur in 1415

At least that’s what Shakespeare had him say, and who am I to dispute Shakespeare? The great bard’s words are quite malleable, one reason they remain so popular today, so I shall make use of them here for something the great writer (and shrewd real estate investor) from Stratford-Upon-Avon never envisioned: The Census Bureau’s annual population estimates.

Most years these aren’t particularly interesting. This year they are, because they show a reversal of some long-standing trends: People moving out of many (though not all) major metros, and people moving into many (though not all) rural areas. These estimates cover the first full year of the pandemic, so are we seeing a COVID-induced, Zoom-era rural renaissance? We don’t know yet. We are rightly cautioned about reading too much into a single year’s estimates for the same reason that we shouldn’t judge a football team by just one game. Sometimes it means something for the whole season; sometimes it doesn’t. Of course, that hasn’t stopped me from writing three columns so far – pointing out the trends, looking at how many people each community would need to have move in to overcome the deficit of deaths outnumbering births, and, finally, looking at which localities have the biggest problems with more people moving out than moving in.

The surprise takeaway from that last one: The population outflows from the coal counties have almost stopped – many coal counties now show more people moving in than out. But the locality with the biggest deficit of more people moving out than moving in is Roanoke. I pointed out that this isn’t new – the out-migration numbers are higher than they have been, but not that much higher. What’s really changed in Roanoke is that the birth rate has fallen, just as it has everywhere.

After that column, I heard from a reader who wanted to know why I focused on the city, rather than the metro area, since businesses tend to look at Metropolitan Statistical Areas. My answer: That’s true, but we govern ourselves by localities, so if there’s a programmatic fix to out-migration, it’s going to have to involve the city government. We shouldn’t try to hide one locality’s big deficit by covering that up in the context of a larger metro area.

Still, it’s a good question: What is the metro-wide tally? So, math time!

Here’s how the Census Bureau estimates the population changes for the localities in the Roanoke MSA for the year ending in June 2021:

Botetourt County: +252

Craig County: +15

Franklin County: +469

Roanoke County: -256

Roanoke: -1,010

Salem: -1

Total population change for Roanoke MSA: -531

Oops! That’s not very helpful – at least not for promotional purposes. It’s actually quite helpful for our understanding of the regional economy because it underscores the importance of our central city. The population growth in the exurbs of Botetourt and Franklin is not enough to make up for the population loss in the city – so it’s in everyone’s interest in the Roanoke metro area to help the city reverse its net out-migration trends. (Or reverse its falling birth rate, but that’s harder to do because that’s subject to forces far beyond the command of any city council.)

These Roanoke metro results naturally make me curious about other metros in our region, so let’s turn now to Lynchburg. Here’s what the Census Bureau says about the localities in the Lynchburg metro:

Amherst County: +11

Appomattox County: +220

Bedford County: +627

Campbell County: -146

Lynchburg: +138

Total population change for Lynchburg metro: +850

Hmmm. So the Roanoke metro shrank last year but the Lynchburg metro grew – thanks mostly to population growth in Bedford County. Maybe some of that Bedford growth is really Roanoke-oriented, but the government doesn’t split localities for MSA purposes – it’s either all or nothing. In any case, it’s clear that most of Bedford’s growth is in the Forest community outside Lynchburg, so let’s not quibble over that. (As I pointed out before, Bedford has far more deaths than births but still has so many new people moving in that they more than make up for that.) For what it’s worth, the Lynchburg metro growth was slower than it was during the past decade, when it averaged a net gain of 1,518 each year, but, unlike the Roanoke metro, it is still growing.

Moving on. We can’t really do this for the New River Valley. Hamilton Lombard, a demographer with the Weldon Cooper Center for Public Service at the University Virginia, advises that data collection in college towns was whacked – my word, not his – during the pandemic and shouldn’t be trusted. The Census Bureau estimated that Montgomery County lost population over the past year. Nobody believes that. We can, at least, look at the other localities in that MSA.

Floyd County: +92

Giles County: -192

Pulaski County: -1

Radford: +410

So the New River Valley probably gained population, we just can’t use these figures to prove it.

The only other officially defined MSA in our region is Kingsport-Bristol-Johnson City, which straddles the state line, which complicates things. The state line also apparently means something: All the localities on the Tennessee side gained population; all the localities on the Virginia side lost population. I’ll have more to say about why that is in a future column but for now let’s just concern ourselves with the Virginia side:

Scott County: -128

Washington County: -204

Bristol: -205

In a previous column, I pointed out how those population losses – particularly in Scott and Washington counties – are driven by deaths outnumbering births in an aging population. Both those localities saw more people move in than move out, so there’s a hopeful trend beneath the surface of what appear at first glance to be negative numbers.

Danville and Pittsylvania County used to be an MSA until being demoted in 2013 to a “micropolitan area” because the core city’s population had fallen below 50,000. Micro, metro, tomahto, tomato – those mean something to somebody but we can still add up the numbers.

Danville: -250

Pittsylvania County: -340

Total population change for Danville micropolitan area: -590

OK, losing population isn’t good but, same as with those Bristol-area localities, I showed in a previous column how Danville and Pittsylvania both now have net in-migration – the number of newcomers moving in is just wiped out by deaths outnumbering births. I wouldn’t worry too much about the overall population numbers here because there are trends afoot that – if properly cultivated and accelerated – can soon turn that minus number into a plus. What we’re seeing there is like a sick patient who is starting to recover. Their temperature is still high, but the medicine is starting to work.

Next up, the Martinsville micropolitan area:

Henry County: -534

Martinsville: +95

Total population change for the Martinsville micropolitan area: -439

This is not quite as hopeful as the Danville area. Martinsville is now seeing enough net in-migration to offset a modest deaths-over-births deficit, but Henry County is seeing a double negative, with deaths outnumbering births and more people moving out than moving in. Martinsville has made steps forward but Henry County is still falling behind, demographically speaking.

The Bluefield micropolitan area is another one that straddles the state line, so it’s Mercer County in West Virginia and Tazewell County in Virginia. Both localities are losing population; Tazewell is down 418, although that’s driven almost entirely by deaths. The in-migration almost balances out the out-migration. The net is -3, which for Tazewell, in historical terms, should be regarded as a great accomplishment.

Our final micropolitan area covers Dickenson County, Norton and Wise County and is officially designated the Big Stone Gap micropolitan area.

Dickenson County: -208

Norton: +9

Wise County: -492

Total population change for Big Stone Gap micropolitan area: -691

Those numbers won’t surprise anyone, although the small population growth in Norton might. Just about every locality is seeing deaths outnumber births. Some localities in Southwest are now seeing net in-migration; these localities aren’t among them.

So what have we learned from this little exercise? Probably not that much from the micropolitans, simply because they’re so small and cover so few localities. However, for the two metros where we have the clearest, most unambiguous numbers – Lynchburg and Roanoke – we have a tale of two cities. One grew slightly, one lost slightly. The Lynchburg MSA is showing population growth in all but one of its localities (Campbell). The Roanoke MSA, though, is more uneven, with its outer localities (Botetourt, Craig, Franklin) growing while its inner ones (Roanoke, Roanoke County and Salem) aren’t.

For me, the big takeaway remains those Roanoke numbers, partly because the city’s out-migration is out of alignment with so many other localities in the region, and partly because they’re large enough that they pull down the whole Roanoke MSA. But we shouldn’t overlook Roanoke County, either. Roanoke County hasn’t lost population in a census since the 1950 count that measured the 1940s. This is just one year, so this trend may go away, but it’s still something to keep an eye on in case this becomes more than just a one-year blip.

For a long time, the dynamics in the Roanoke Valley was the city versus the county. That dynamic also held back the region for a long time. Here we see both localities with the same problem: population loss. It’s caused by different things – in the city, there’s a small births-over-deaths number and a high out-migration number; in the county, there’s a large deaths-over-births number and a small in-migration number. Roanoke County is simply aging out – and dying out. The city is seeing people move out, but it’s not as if they’re all moving to the county. (The city’s net out-migration was 1,169, the county’s net in-migration was 192.) Different problems, different causes, but ultimately the same solution: Both Roanoke and Roanoke County need more people, ideally young adults, moving in. Otherwise, they will both continue to lose population – and, depending on the numbers, so might the whole Roanoke metro.

Again, I must caution that these out-migration trends aren’t new – we’re just seeing it now in the same way that we saw the foundations of old buildings emerge the year that Carvins Cove, the city’s reservoir, suffered a drought. That means we can’t say “oh, this is happening because the crime rate is up” or “this is because city council did such-and-so” – because this is a more structural problem than some sudden response to some recent event.

There’s also some important context: Many metro areas are seeing out-migration exceed in-migration, so Roanoke’s not unique. I’ll look more deeply at those numbers in tomorrow’s column.

Yancey is editor of Cardinal News. His opinions are his own. You can reach him at