44% of the Roanoke Valley’s population growth is in Salem. Roanoke ranks last for attracting newcomers.

We’ve collected all our work on Virginia’s changing demography in one place; find it here.

Roanoke voters this fall will pick three members of the city council. The seven candidates seeking those seats are talking about the most contentious issues in the city: whether Roanoke needs a casino, at what level schools should be funded, what can be done to build more housing.

Next door, Salem is electing two members of its council, and much of the conversation there also revolves around housing.

What none of the candidates in either city are talking about are the demographics that are driving some of these issues. Today, in hopes of sparking that discussion, I will examine how the demographics of the two neighboring cities have changed and how that contributes to some of the stresses both are now feeling. Unless otherwise noted, all these numbers come from the U.S. Census Bureau, which the Charlottesville-based demographer Hamilton Lombard has helped me find. Let’s walk through them.

44% of the valley’s population growth is in Salem, almost none in Roanoke

The most recent population estimates from the Weldon Cooper Center for Public Service at the University of Virginia show that from 2020 to 2025, Roanoke’s population grew by 75 people — a rate of 0.1% — to 100,086. That’s a slight improvement over previous trends, which had shown the city’s population shrinking slightly. All things are relative: Roanoke has turned around a small decline. We can have a robust debate about how much population growth is too much, but few would dispute that losing population is not a good thing.

Roanoke’s slow population growth rate has been pretty much similar to Roanoke County, whose population grew by 0.3% (an increase of 331 people) and Botetourt County, which saw an increase of 0.7% (228 people). The exception was Salem, which saw its population tick up by 2.1%. That’s an increase of 522 people. That means Salem has added more people than Roanoke and Roanoke County combined. Put another way, 44% of the population growth in the Roanoke Valley has been in Salem. There’s a reason why housing is a hot topic in Salem.

When people move to the Roanoke Valley, Roanoke is their last choice

The city may be the first choice for some new residents, but overall, it’s not. Here’s the math:

Although Roanoke’s total population has been essentially flat, there’s been lots of movement beneath the surface — people are born, people die, people move in, people move out. Just because the surface numbers look stable doesn’t mean the numbers beneath them are.

In Roanoke, Roanoke County, Salem and Botetourt County, deaths outnumber births. The reason the population in those localities is growing, however slightly, is more that people are moving in than anything else.

In Roanoke during the first half of the decade, deaths outnumbered births by 913 — or, as the demographers record it, -913. That, by itself, is not an unusual number. Roanoke County’s population is aging out faster; in the county, deaths outnumbered births by 1,920. This is a consequence of both an aging population and a declining birth rate. In most localities in the state outside Northern Virginia, deaths outnumber births.

Now let’s look at the net migration that’s offsetting it. The good news is that all the localities in the Roanoke Valley have net in-migration — more people coming in than going out.

Here are the numbers:

Roanoke County: 2,251
Salem: 1,393
Botetourt County: 1,094
Roanoke 988

This is the basis for me saying that Roanoke is the last choice for those moving into the valley. On a percentage basis, only 17.2% of the net migration into the valley is to the city.

The census data can’t tell us why that is — whether it’s a matter of choice or housing availability or something else. The only real way to know is to ask all the people moving in. However, if the reason is a matter of housing availability, then that’s something the council can address through zoning and other means, although many proposed developments in the city (such as developing the wooded Evans Spring property) have been controversial. It’s perfectly valid, of course, for the city not to develop that property or any other if the council chooses; there are simply trade-offs to everything.

Here’s one way for these communities to think about housing and migration: How many people does each one need to attract to offset deaths outnumbering births? Here are those numbers:

Deaths vs. births
Roanoke County: -1,920
Roanoke: -913
Salem: -871
Botetourt County: -866

You might be wondering: If in-migration perfectly balances the death deficit, why do we need any new housing? Doesn’t each death open up a residence? Not necessarily. Imagine a house that once was home to four people, two adults and two children. The children have now grown up and moved out, so that’s -2 on migration. Then one of the adults dies, and we’re -1 on deaths. To balance those out, we need three new people to move in — but the house is still occupied by the lone survivor. That’s why even communities losing population need new housing.

Roanoke’s relatively low net in-migration rate isn’t simply a function of being a city

It’s certainly true that suburban localities have more room to build new housing; just look at all the construction in the Daleville area of Botetourt County. It’s also true that some central cities have trouble attracting new residents. Roanoke is certainly in better shape than Portsmouth, which has net migration of -1,240, meaning 1,240 more people moved out than in. Or Newport News, which records -6,665. Roanoke’s 988 net migration looks pretty darned good compared to them.

However, for other comparisons, note that Martinsville has had a net in-migration of 895, Lynchburg has had 2,006 and Danville 2,261. Danville is definitely living up to its moniker of “the comeback city.” Again, we come back to a policy choice from the city council: How much in-migration does Roanoke want? And what can it do to achieve that? Maybe the current level is sufficient, maybe it’s not. That’s for voters — and their council members — to decide.

Roanoke’s child population is dropping, Salem’s is growing

Now let’s delve deeper. Roanoke’s population, while staying essentially flat on the surface, is changing in very specific ways underneath. Here’s one big way: It’s losing children.

From 2020 to 2025, Roanoke’s population of those from birth to 14 dropped by 3.1%. That was generally in keeping with other localities in the western part of the state:

By itself, this shrinkage is not particularly unusual; we are in an era of declining birth rates. We see a lot of localities that are seeing their total population grow, but the number of children fall.

What is unusual is how Salem pops out on this map. Salem is the only locality in the Roanoke Valley (and one of the few in the western part of Virginia) that’s seeing an increase in its 0-14 population. A decade ago, Salem’s 0-14 population was shrinking. Now it’s growing — at a rate of 8.2% during the first half of this decade.

Here’s what Salem voters (and the officials they elect) should take note of: Many of these kids haven’t hit the school system yet.

You’ll see from this chart that the number of Salem residents 10-14 has declined, while those 5-9 have stayed about the same. The growth has come in those ages 0-4; that age group in Salem has grown by 28.8%, which puts it in the top six in the state for growth in that age group on a percentage basis. Salem schools appear likely to see an influx of new students in the years ahead. Is the city prepared?

Roanoke is attracting young adults, but the rate is slowing

The Census Bureau defines “young adult” as ages 25-44. A decade ago, most Virginia localities outside the urban crescent were losing people in this age group, although Roanoke and Salem stood out as the rare exceptions.

Now, here’s what’s happening in the first half of this decade.

This is a major turnaround. We’re now seeing a youth movement (those 44 years old will appreciate being called young) into much of rural Virginia. (I’ll deal with this more in-depth in a future column.) Roanoke, though, is going the opposite direction. It’s still attracting young adults but at a slower rate: 1.2%.

One reason for that slowdown may simply be that young adults who once would have moved to Roanoke are now moving (or staying) elsewhere. Botetourt County and Roanoke County were losing this age group a decade ago; now they’re growing. Meanwhile, Salem, which was attracting this age group before, is now attracting those people at a faster rate of 9.3%.

When we look more closely at five-year age increments, we see that in Roanoke the growth in the 25-44 age range is taking place entirely from those 35-44. Salem is seeing fairly consistent growth throughout the 25-44 age range, but Roanoke’s is entirely on the older side, which suggests that those 25-34 are actively opting against the city. Why? The data can’t tell us. We just know that among those 25-29, every other locality in the Roanoke Valley is seeing growth, but the city is seeing a decline. Among those 30-34, both Salem and Roanoke County are seeing growth, but the city and Botetourt County are seeing declines. Something is driving that.

Here’s one way to look at this: A decade ago, Roanoke and Salem had a near-monopoly on growth in the 25-44 age group in the region. Now they don’t; there’s simply more competition. If Roanoke wants to attract more residents in this age group, it will need to work harder to attract them. What does that mean in practice? That’s a good policy debate.

Roanoke and Salem are both losing residents 45-64

The Census Bureau calls this age group “middle-aged.” However, with many people delaying childbirth into their 30s, some of the people on the younger end of the 45-64 bracket are likely parents of children in that 0-14 age group. When those parents move, so do the kids.

You can see from this map that this is not unusual; it’s a widespread trend.

Some of this decline is simply a matter of a smaller cohort to begin with — the first wave of the declining childbirth rates decades ago. Some of it may be people maxing out their careers in the region and needing to move elsewhere to advance professionally. Some of it may be people retiring early and also moving elsewhere. All that requires more analysis than the data I have at hand.

Still, the point is, this is the age group both cities are losing. That has ramifications in all these communities — that’s the age group that is often involved in civic activities. With fewer people in those age groups, those civic responsibilities are either pushed upward to older residents or downward to younger residents — or maybe simply don’t get done.

Almost everywhere is growing older; Roanoke and Salem aren’t exceptions

We’re an aging society. Almost every locality is growing older. The only thing notable here with Roanoke and Salem is that their rate of aging is slowing down. It’s localities in the urban crescent that are aging faster.

The policy choices on the table

The first question voters (and their elected officials) need to answer is what they want to do about the volume of deaths over births. There’s probably nothing local governments can do about deaths. They can indirectly help increase the number of births by increasing the number of young adults — figuring out how to do that gets into other issues, such as housing, economic development and various quality-of-life amenities.

The next question to deal with is the one I noted above: How much in-migration do these communities want? If they want to just stay even, then they know the targets. If they want to grow their population, then they need to exceed those targets.

As for housing, how much do Roanoke and Salem need to create? What kind and where?

As for economic development, that’s a regional question: A jobs announcement in, say, Botetourt County might lead to jobs for people in each locality — and housing demand in places other than Botetourt.

Housing is something local governments can more directly encourage through zoning decisions. Attracting jobs is trickier; simply rezoning a piece of land won’t necessarily persuade a company to locate there. Much of the land in industrial parks sits vacant for decades; that’s simply the nature of the business — voters need more patience than they often have.

Those policy questions are for voters and opinions may vary — but these are the numbers they need to deal with, whether they like them or not.

In the weeks ahead, I’ll provide similar demographic analyses for other localities in our coverage area. Among them: Danville, Lynchburg, Martinsville and any others that stand out for having unusual trends. If you have a request, let me know, and I’ll see what I can do.

We have more politics each week in West of the Capital, our weekly political newsletter that comes out on Friday. This week, I’ll take a look at the latest early voting trends across Virginia plus invite readers to vote on their favorite precinct names. You can sign up here: