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

Last week the U.S. Census Bureau released its annual population estimates – separate from the every-10-years headcount – and those estimates were far more interesting than you’d think a bunch of numbers could ever be.

The estimates (which aren’t guesswork; there is data that goes into them) were significant because they give us our first glimpse into how the first full year of the pandemic has changed a lot of American trends.

First, the estimates showed that 73% of American counties saw deaths exceed births – up from 55.5% the year before.

Second, the estimates showed that 65% of American counties saw more people moving in than moving out.

Put another way, we saw certain trends accelerate. Birth rates are falling, which means even without COVID deaths, the deaths-over-births count is widening. That suggests that the nation’s population may someday fall, unless there’s more immigration. Meanwhile, for at least one year, there really does seem to be a Zoom-era migration out of some (but certainly not all) big cities and into rural areas. Those localities may still be losing population, though, because all those deaths outnumber both births and newcomers moving in.

I dealt with that in Monday’s column. Today I’ll look at those numbers a different way. Beware: Math ahead, but also some insights into some of the trends – some hopeful, some not – that are reshaping Southwest and Southside.

On Monday, to look at where trends were accelerating, I compared the 2021 estimates with the pre-pandemic 2019 estimates so we saw both how the margin of deaths over births was increasing, but also how domestic migration was increasing. Today, I’ll look strictly at the 2021 data, which covers July 2020 through the end of June 2021, so encompasses a full year of the pandemic. I’m also, for the sake of simplicity, only going to look at the two main parts of the 2021 data: deaths versus births (natural increase or decrease) and net domestic migration (whether more people are moving in or moving out). That means I’m leaving out the figures on international migration for several reasons. First, we have relatively little international migration in Southwest and Southside (a missed economic opportunity for us, by the way). Second, those numbers really are estimates, according to Hamilton Lombard, a demographer with the Weldon Cooper Center for Public Service at the University of Virginia. By contrast, the domestic migration figures come from the Internal Revenue Service, which keeps track of where you’re filing your income taxes from. So for our rough, back-of-the-envelope calculations here today, those are the figures we’ll concentrate on.

Now, before we get into the data, I must issue some caveats. First, Lombard cautions that we shouldn’t read too much into one-year data, in much the same way that we shouldn’t look at a single football game and draw too many conclusions about how the team will do over the full season. Sometimes upsets happen. Or, to quote the country-rock singer Ray Wylie Hubbard – who will appear in May at Epperly Mill in Floyd County – “Everybody turns a bad trick now and then.” Second, we don’t know yet whether these numbers are abnormal – a one-year blip due to the pandemic – or represent some kind of “new normal.” Lombard seems to think they might. Certainly death rates have been elevated due to COVID, but we also have an aging baby boom generation getting ready to, as Shakespeare once said, shuffle off the mortal coil, so we can’t necessarily say that deaths will go down much post-COVID. (I want bonus points for working both the iconic Shakespeare and the cult figure Hubbard into the same column on demographics; gotta liven up the numbers somehow.) Lombard also says there were some hints before the pandemic that people might be cooling on the idea of living in major metros; now we’re definitely seeing more hints of a migration during the first full year of the pandemic – the big question is whether that will continue?

That brings us to the question I want to explore today: What would it take for each locality in Southwest and Southside to gain population? Before we can talk about policies and programs, let’s see what’s involved numerically – how close or far away is that goal?

Let’s organize the discussion this way:

Localities where births outnumbered deaths in 2021 (and by 2021, I mean the year ending June 2021)

This is easy because there are only five. This includes the two biggest metros (Lynchburg and Roanoke), two small college cities (Lexington and Radford), and one county that’s just outside one of those big metros but close enough to be a commuter county (Appomattox County). If you’re wondering why Blacksburg isn’t on the list, blame Virginia’s system of local government. Blacksburg is a town so its statistics are counted as part of Montgomery County.

Three of those five were the only localities in our region with a double plus. They saw population growth in two ways: from births outnumbering deaths and from more people moving in than out.

Appomattox County: 4 more births than deaths; 212 more people moved in than moved out

Lexington: 45 more births than deaths; 95 more people moved in than moved out

Radford: seven more births than deaths; 404 more people moved in than out

Appomattox was one of the few counties in Southside that gained population in the last census, so the strong in-migration here isn’t out of context. Appomattox is increasingly functioning as an exurb of Lynchburg. (It’s also home to the cutting-edge Wolfbane Productions, which Amy Trent recently wrote about.)

Lombard says we should be cautious about data from college towns because without an actual headcount we can’t be exactly sure whether these figures are capturing students or not. Still, if these figures are right, then at least for 2021, Lexington and Radford were doing things right. (Assuming they want population growth, of course. You can obviously have too much, but losing population is also tied with a failing economy – fewer customers for local businesses, all that – so I’m working off the assumption that at least a little population growth is a good thing.)

The other two localities were in a single-plus category – more births than deaths, but more people moving out than moving in.

Lynchburg saw 178 more births than deaths, but then saw net out-migration – 116 more people moved out than in. The net result: a slight population increase.

Roanoke is more curious – and worrisome. Roanoke is in the odd position of having more births than deaths – something that’s unusual – yet losing population because more people moved out than in – something that’s also unusual.

The city saw 34 more births than deaths – but then 1,169 more people moving out than in. That seems a shocking net out-migration for a city of 100,011 and runs counter to the population growth of 3.07% that Roanoke saw over the past decade in the census. Why did so many people decide in 2021 to move out of Roanoke? This is something that merits a lot more inquiry. If these estimates are right, then one year alone just wiped out one-third of the population growth from the past decade. This seems very odd to me because, even prior to the pandemic, Roanoke was showing up as a center for remote workers – and a LinkedIn survey last September showed Roanoke with one of the highest national averages for workers applying for remote jobs.

Lombard points out, though, that eight of the 10 years during the past decade, Roanoke had net out-migration, with the biggest deficit being -1,076 in 2016. This number is higher than that, but not that much higher. The only reason Roanoke gained population during the past decade isn’t because more people moved in – they didn’t – but because the city saw more births than deaths. Now the birth rate is falling (it’s falling almost everywhere), so just as a receding lake reveals things we hadn’t seen before, the slowing birth rate exposes Roanoke’s out-migration. In this way, Roanoke is like many major metros – the pandemic has accelerated an out-migration. Unfortunately, this isn’t the way Roanoke wanted to emulate some of those bigger cities. My advice to the Roanoke City Council and those who would like to be on the city council in this fall’s election: Devote some serious discussion to how to reverse the city’s net out-migration because these trends suggest the birth rate will no longer keep covering that up. We can’t do much about the death rate or even the birth rate. (Russia has tried to raise its birth rate as a matter of national policy and failed.) But localities can do things to make themselves more attractive to both the newcomers they’d like to attract and the current residents they’d like to retain.

All the other localities in Southwest and Southside saw deaths outnumber births, so we need some logical way to organize these. Here’s how we’ll do it.

Localities where deaths outnumbered births but net in-migration made up the difference

LocalityDeaths over birthsNet gain of people moving in vs. people moving out
Amherst County-158159
Bedford County-4301,042
Botetourt County-249499
Brunswick County-107238
Buckingham County-69197
Cumberland County-4469
Floyd County-36126
Franklin County-311776
Grayson County-133190
Greensville County-3477
Patrick County-221266
Prince Edward County-4196
Rockbridge County-175206

My first takeaway here: This is a lot of localities in Southwest and Southside that, at least for one statistical year, saw a population growth.

Some of these aren’t surprising: Bedford County has been gaining population for decades (the last time it lost population was during the 1960s, which showed up in the 1970 census). Bedford has three drivers of growth: Roanoke to the west, Lynchburg to the east and Smith Mountain Lake to the south.

Others are more interesting. Of the localities on this list, only three – Botetourt, Floyd and Rockbridge – gained population in the 2020 census. So that means the other 10 localities showed population growth during 2021 when they didn’t before.

In some cases, this is a return to form. Franklin County had been gaining population since the 1950s, often by double-digit rates, until it all came to a halt in the past decade – a combination of the 2008 recession slowing the housing boom at Smith Mountain Lake, and aging demographics finally catching up with the county. Now, once again, the county is gaining enough newcomers to outnumber those who are dying.

In other cases, this is a turnaround – however brief it might be – of historic proportions. Martinsville has been losing population in every census since the 1970s. Grayson County has lost population in three of the last four census counts and six of the last eight, going back to 1950, which measured the 1940s outflow. Brunswick County has lost population in eight of the past 10 census counts, going back to 1930, which measured the 1920s. We often think of the coal counties as classic examples of places that peaked in population long ago, but Brunswick’s peak predates theirs: Its peak population came in 1930 and the county has generally been shrinking ever since.

That’s why this is potentially a big deal – again, if these estimates are right and if these trends continue. Still, for these localities, this is happy news for places that haven’t seen a lot of happy news.

Now we come to another category:

Localities where deaths outnumbered births but net in-migration almost made up the difference.

“Almost” is a subjective term, so some might quibble about where I drew the line.

LocalityDeaths over birthsNet gain of people moving in vs. people moving out
Craig County-4226
Buena Vista-221
Highland County-2111
Lunenburg County-10484
Mecklenburg County-296236
Nottoway County-4218
Pulaski County-250246

Of these localities, Pulaski County is the closest: It needed just four more people moving in during 2021 to balance out deaths over births. If five moved in, it would have shown population growth. Covington just needed eight, Highland County 10. Granted, 10 people is a bigger percentage in Highland than it is elsewhere but in terms of sheer headcount, all these localities are close. Mecklenburg County was the furthest away from breaking even – it needed 60 more people. Fifteen families of four would have done it, so all these seem potentially achievable numbers.

Now we move on to the more difficult cases.

LocalityDeaths over birthsNet gain of people moving in vs. people moving out
Campbell County-22034
Carroll County-243145
Halifax County-23867
Lee County-210109
Pittsylvania County-42976
Roanoke County-505192
Scott County-249116
Washington County-447232
Wythe County-18799

Some localities on this list may not want to be here, others may consider it an improvement from where they were.

The one that surprises me most is Campbell County. It’s right outside Lynchburg. Why isn’t it seeing more in-migration the way Bedford and Appomattox are? Roanoke County reflects the aging of its suburban population, and perhaps a preference by many young adults for downtown living. (Just not enough, obviously, to counter the huge out-migration in the city discussed above.)

Danville and Pittsylvania County seem actually hopeful, particularly with the economic transformation taking place in what the city now likes to call “the comeback city.” Pittsylvania has seen out-migration outnumber in-migration for the past five years, Danville six years in a row. Both have seen out-migration outnumber in-migration for eight of the past 10 years. If these figures mark the start of a new trend, they lay down a marker of just how many people the two localities need to have moving in to halt overall population decline due to deaths over births – in Danville, a net in-migration of 276; in Pittsylvania, a net in-migration of 353. Of course, future years may not have the same number of deaths – they might have more. Those localities should be doing some demographic number-crunching so they can set some realistic metrics.

For localities in far Southwest such as Lee County, Scott County and Washington County, these figures also seem quite hopeful. Washington has had more people move out than move in for six years in a row; now that’s changed. For Scott, seven of the past 10. For Lee County, nine years in a row.

It’s hard to overcome the big death rates there – all those localities have some of the oldest median ages in the state so they’re going to have to deal with a lot of deaths. But reversing negative outflows seems significant. What we don’t know yet is just who these newcomers are.

“If they tend to be younger,” Lombard says, “that could help chip away at the birth/death imbalance nearly every rural county has.” Seems to me these localities have something to be encouraged about.

Now we come to the final, more difficult, category:

Localities where deaths outnumbered births – and more people are still moving out than moving in.

We had three localities at the beginning of this list in a double-plus situation; these 17 are all in the double-minus range.

LocalityDeaths over birthsNet loss of people moving out vs. people moving in
Alleghany County-164-13
Bath County-21-61
Bland County-49-25
Buchanan County-228-176
Dickenson County-152-106
Giles County-125-69
Henry County-500-40
Montgomery County-110-1,075
Russell County-173-6
Smyth County-335-73
Tazewell County-427-3
Wise County-255-118

Notice that almost all these localities are in Southwest Virginia, particularly in or near the coal counties. Montgomery County is the oddity – it’s previously had more births than deaths and more people moving in than out. The former change can be attributed to the overall decline of the birth rate but the latter is hard to fathom, although Lombard says the pandemic has messed up a lot of data analysis for college towns. We should just pass on this one and come back to it in a year.

Some of these obviously aren’t that bad. Alleghany County,Galax, Norton, Russell County and Tazewell County are pretty close to breaking even on migration. All those are big changes from previous outflows which were definitely in the negative range. Even Buchanan County‘s outflows are slower than they have been for six of the past nine years. Sometimes before you can stop the car you have to slow it down. Maybe that’s happening. Maybe it’s not. The question is whether Buchanan can turn losing a net of 176 people a year into a net gain of 228 to balance out deaths. That’s the challenge for all the localities but, given the overall trends, perhaps an opportunity, as well.

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