The following article by Ruy Teixeira, author of The Optimistic Leftistand other works of political analysis, is cross-posted from his blog:

I know I write a lot about how key the Rustbelt will be in this election, particularly the states of Michigan, Wisconsin and Pennsylvania. I stand by that view for what I believe are sound empirical reasons. But there are also sound empirical reasons for thinking about the Democrats’ future beyond this election as primarily lying in a different part of the country.

As Ron Brownstein notes:

“For Democrats, the Sun Belt imperative is growing more urgent.

While most in the party are preoccupied with winning back the three Rust Belt states that tipped the 2016 election to Donald Trump, both people and political power are continuing to migrate inexorably from that region to the younger and more diverse states in the Southeast and Southwest.

This sustained population shift reinforces the consequences of Trump’s political repositioning of the Republican Party. Trump has targeted his polarizing message and agenda heavily toward the priorities of the older and non-college-educated white voters who still dominate most of the Rust Belt. That will make it tough for Democrats to rely on those states, particularly in presidential races, as much as they did during the 1990s and earlier this century.

In the near future, then, Democrats will likely need to offset any Republican gains in the Rust Belt by winning more elections in Sun Belt states, which are adding more of the diverse, white-collar, and urbanized voters at the core of the modern Democratic coalition. Through the coming decade and beyond, the crucial variable that could tilt the national balance of power between the parties may be whether Democrats can leverage those demographic advantages in the Sun Belt to break the hold Republicans have enjoyed on most of the region since at least the 1970s.”

In aid of thinking about trends in these states and possibilities for the future, I offer the coverage in my Path to 270 in 2020 report of these states.

The Southwest

The Southwest includes five states that could be in play between the Democratic nominee and Trump:

• Texas: 38 electoral votes
• Arizona: 11 electoral votes
• Colorado: nine electoral votes
• Nevada: six electoral votes
• New Mexico: five electoral votes

Together, these five Southwestern target states have 69 electoral votes. In 2016, Trump carried Texas and Arizona, and Clinton took Colorado, Nevada, and New Mexico. While Texas is still likely in the Republican camp for this cycle, despite recent positive trends, Arizona is a stronger—and important—possibility for Democrats. Arizona could put the Democrats over 270 if the their candidate took Michigan and Pennsylvania but failed to take Wisconsin.

The GOP strategy will focus on safeguarding Arizona and Texas and trying to pick off one or two other Southwestern swing states as insurance against Rust Belt losses. The Trump campaign has publicly mentioned New Mexico and Nevada as targets.

These Southwestern states are all fast growing relative to the national average. They also have relatively large nonwhite populations, especially compared with the Midwest/Rust Belt states. Overall, these Southwestern states present a demographic profile and growth dynamic that is more favorable for the Democratic nominee than the Midwest/Rust Belt swing region, where the heavily white populations and slow pace of demographic change are relatively advantageous for the GOP. Below, we provide a detailed discussion of these states in descending order of electoral votes.

Texas: 38 electoral votes

Trump won Texas by 9 points in 2016, a significant drop from Romney’s 16-point victory four years earlier. Despite this enticing improvement for the Democrats, it should be emphasized that Republicans have carried the state since 1976.

Democrats also had some successes in Texas in 2018. They lost the House popular vote by less than 4 points—a big advance for them in the state—and flipped two GOP-held House seats. Moreover, since that election, no fewer than six GOP House incumbents have announced their retirements, creating further possibilities for the Democrats. The Democrats also flipped 14 state legislative seats from the GOP and broke their supermajority in the upper chamber. Finally, while Republicans won handily by a double-digit margin in the governor’s race, Democrats made the race against incumbent GOP Sen. Ted Cruz far closer than almost anyone thought possible; Democrat Beto O’Rourke wound up losing by less than 3 points.

The Democratic candidate in 2020 will seek to build on these trends. But of course, Democrats would need quite a swing relative to 2016 to succeed in flipping the state. Trump only needs to come close to the voting patterns he benefited from in 2016 to once again carry the state. Currently, he has only a modest +3 net approval rating in the state, so that is not something he can take for granted.

Texas has a huge nonwhite population, though it is somewhat less represented among actual voters. In 2016, nonwhites made up 39 percent of voters in the state in 2016—13 percent Black; 21 percent Hispanic; and 5 percent Asian/other races. Blacks and Hispanics supported Clinton by 76 points and 26 points, respectively. Asians/other races, however, supported Trump by 13 points. Texas white college graduates (27 percent of voters) also supported Trump by 20 points, 57 percent to 37 percent, while the largest group—white noncollege voters (34 percent)—backed him by a whopping 55 points, 76 percent to 21 percent.

Our estimates indicate that white noncollege eligible voters in 2020 should decline by 2.5 points relative to 2016, while white college graduates should also decline, though only very slightly. Black eligible voters should remain roughly stable, while Hispanics should increase by more than 2 points as a share of eligible voters, and Asians/other race will go up by half a point. On net, these changes favor the Democrats and will put a modest dent—1.6 points—in the GOP advantage in the state if voting patterns by group do not change in 2020.

As Trump’s massive lead among white noncollege voters suggests, if he can maintain or come close to his support among this group in 2020, he will most likely win Texas. Even a shift of 10 margin points against Trump among white college graduates, continuing a recent pro-Democratic trend, would still leave him with a 5-point lead in the state.

For the Democratic candidate, a winning coalition would have to include not only this big white college swing but also a large (15 margin points or so) pro-Democratic swing among Hispanics, Asians, and voters of other races, combined with increased nonwhite turnout overall. But even with these favorable changes, the Democratic candidate probably needs to reduce at least slightly the massive deficit among the white noncollege group. All in all, one would still have to favor Trump to take the state, but certainly the trends from 2018 onward suggest that Democrats may be able to take advantage of some of these pro-Democratic changes and that the state could be quite competitive in 2020.

Arizona: 11 electoral votes

Trump won Arizona by just 3.5 points in 2016, a substantial drop from Romney’s 9-point margin in 2012. Republicans have carried the state since 1996, but the 2016 result has given Democrats hope they can carry the state in 2020 for the first time in decades.

Democrats reduced this deficit further in Arizona in 2018. They won the House popular vote by just less than 2 points and flipped a GOP House seat. The Democrats also flipped four state legislative seats from the GOP. Finally, and most importantly, they flipped one of Arizona’s GOP-held Senate seats, as Democrat Kyrsten Sinema defeated Republican Martha McSally by 2 points. In the governor’s race, however, the Republican candidate soundly beat the Democrat by double digits.

The Democratic candidate in 2020 will have a lot of upward trends to build on to turn 2016’s close loss into a close victory in 2020. As for Trump, he will need to hold the line from 2016 and make voting patterns in 2020 as much like the previous election’s as possible. Adding to that challenge, he currently has a negative net approval rating in the state of -5.

Arizona has a substantial nonwhite population, though, as with Texas, it is somewhat less represented among actual voters. In 2016, nonwhites made up 27 percent of voters in the state in 2016—17 percent Hispanic; 6 percent Asian/other races (a group that includes Native Americans); and just 4 percent Black. Hispanics supported Clinton by 36 points; Blacks by 52 points; and Asians/other races by 8 points. Arizona’s white college graduates (30 percent of voters) supported Trump only narrowly, by 47 percent to 46 percent, while noncollege whites, 44 percent of voters, backed him by 27 points, 60 percent to 33 percent.

We expect white noncollege eligible voters in 2020 to decline by almost 3 points relative to 2016, while white college-graduate eligible voters should remain stable. Black eligible voters should also remain roughly stable, while Hispanic voters should increase by more than 2 points and Asians/other races by half a point. These changes in the underlying demographic structure of the electorate are enough to knock a point off Trump’s advantage in 2020, even if voting patterns from 2016 remain in force.

Given the narrowness of Trump’s victory in 2016 and the projected deterioration in his margin from demographic change, Trump needs, at minimum, to hold his 2016 levels of support from various demographic groups. His most effective safeguard against losing the state would be to increase his support among his friendliest group, white noncollege voters. A 10-point margin shift in his favor among these voters would take his projected advantage in the state up to 7 points, all other voting patterns remaining the same.

For the Democratic candidate, a winning coalition could be assembled in several different ways. A 10-point pro-Democratic margin shift among white college graduates (going from -1 points to +9 points) would be enough to generate a half-point victory in the state. A 15-point pro-Democratic swing among Hispanics, Asians, and voters of other races, would be even more effective, taking the victory margin over a point. And a 10-point pro-Democratic margin shift among white noncollege voters would take the Democratic candidate’s advantage to just less than 2 points. Given that a number of trends seen in 2018 were consistent with these possible changes and that Trump’s margin in 2016 was already so thin, Trump may have difficulty holding the state in 2020.

Colorado: 9 electoral votes

Clinton won Colorado by 5 points in 2016. This was essentially the same margin with which Obama carried the state in 2012.

Democrats also had several successes in the 2018 election in Colorado. They carried the House popular vote by a strong 11 points and flipped a GOP-held House seat. They also flipped eight state legislative seats and took control of the upper chamber, thereby giving them a trifecta of control in the state. In addition, Jared Polis held the governorship for the Democrats with an easy 11-point victory.

The Democratic candidate in 2020 obviously has a lot of upward trends to build on in the state. Trump’s path looks distinctly more difficult, especially considering his current negative net approval rating of -16.

Nonwhites made up 19 percent of Colorado voters in 2016, most of whom were Hispanic, at 12 percent, with Blacks and Asians/other races at 4 percent each. Blacks supported Clinton by 60 points; Hispanics by 32 points; and Asians/other races by 13 points. In addition, white college graduates, an imposing 40 percent of voters, backed Clinton by 52 percent to 40 percent. The relative bright spot for Trump was white noncollege voters—41 percent of the voting electorate—who favored him by 15 points.

Blacks and white college voters should remain stable as a share of eligible voters in 2020, while Hispanics will increase by 1.6 points and Asians/other races by half a point. The sole declining group will be white noncollege voters, who are projected to decline by slightly more than 2 points. All these changes favor the Democrats.

The logical strategic choice for Trump would be to pump up his 15-point margin among white noncollege voters from 2016. However, even a 10-point margin shift in Trump’s direction among this demographic would leave him a point and a half behind in the state, all else remaining the same. To succeed, he would probably need to also reduce his deficit among white college voters significantly, which is a much more difficult target.

The Democratic candidate could expand Clinton’s margin by half a point by holding Democratic margins at their 2016 levels, due to underlying demographic changes in the eligible electorate. If they expanded their already strong lead among white college graduates by 10 margin points, that would add 4 points to their advantage. Furthering their lead by 15 points among Hispanics, Asians, and voters of other races would increase their lead by around 2.5 points.

Nevada: 6 electoral votes

Clinton won Nevada by about 2.5 points in 2016, down from a Democratic margin of 7 points in 2012 and more than 12 points in 2008. This trend line has earned Nevada a place on the Trump campaign’s short target list of states that Clinton carried in 2016.

But the other side of the coin here is how well the Democrats did in 2018. They won the House popular vote by 6 points and flipped two House seats. They also flipped four state legislative seats, including attaining a supermajority in the upper chamber. Most importantly, Democrat Steve Sisolak took the governorship for the Democrats by 4 points, giving them a trifecta in state government. To top it off, Democrat Jacky Rosen defeated Republican Senatorial incumbent Dean Heller by 5 points, giving Democrats control of both of Nevada’s Senate seats.

The basis would certainly appear to be there for the Democratic candidate in 2020 to defend the state against the Trump campaign. The closeness of the 2016 race will give his campaign hope, but Trump’s extremely challenging -21 negative net job approval rating in the state underscores the difficulty of his quest.

Nonwhites made up a large 35 percent of Nevada voters in 2016, dominated by Hispanics at 16 percent, but also including Blacks at 9 percent and Asians/other races at 10 percent. Clinton’s margins among these groups were, respectively, 29 points, 56 points, and 2 points. White college graduates, 22 percent of voters, also went for Clinton, though just barely, by a point, 47 percent to 46 percent. But the dominant group, white noncollege voters, 43 percent of the total electorate, gave Trump an 18-point lead (56 percent to 38 percent), thereby bringing him close in the state.

Looking forward to 2020, we project an unusually large decline in white noncollege eligible voters of more than 3 points. White college-educated eligible voters should also decline, though only modestly, by 0.3 points. Black eligible voters should remain roughly stable while Hispanics should increase more than 2 points Asians/other races by around a point. As is typical, these changes in the underlying structure of the eligible electorate should favor the Democrats

Trump’s target in the state for 2020 is clear: enhance his 18-point margin among white noncollege voters from 2016. Increasing this margin by 10 points would give him a narrow victory in the state of less than a point, all else remaining equal. A similar result would obtain if Trump manages the more difficult task of increasing his performance among Hispanics, Asians, and voters of other races by 15 margin points.

The Democratic candidate could expand Clinton’s margin in the state by a point by holding Democratic margins at their 2016 levels, due to underlying demographic changes. Expanding their modest lead among white college graduates by 10 margin points would add 2 points to their advantage, but increasing their 2016 lead by 15 points among Hispanics, Asians, and voters of other races would boost their projected margin by more than 4 points.

New Mexico: 5 electoral votes

Clinton won the state by more than 8 points in 2016, though this margin is down from Democratic margins of 10 points and 15 points, respectively, in 2012 and 2008. Perhaps this trend line is why the Trump campaign says it will attempt to flip the state in 2020.

The 2018 election does not provide much basis for this GOP aspiration. The Democrats overwhelmingly won the House popular vote by 21 points and flipped the GOP’s one House seat in the state. The Democrats also flipped a net of eight state legislative seats and, most importantly, Democrat Michelle Lujan Grisham won the governorship by more than 14 points, flipping that office and thereby giving Democrats trifecta control of state government.

These results do not inspire confidence in the plausibility of Trump flipping the state in 2020. Nor does Trump’s -10 negative net job approval rating in the state, though at least it is better than his Nevada rating.

Nonwhites made up a whopping 47 percent of New Mexico voters in 2016, the highest in the country, outside of Hawaii. These nonwhite voters were dominated by Hispanics at 36 percent of all voters, followed by Asians/other races (a group that includes Native Americans) at 9 percent, and Blacks at 2 percent. Clinton carried all these groups by, respectively, 26 points, 7 points, and 51 points. White college graduates, 25 percent of voters, also went strongly for Clinton by 17 points, 53 percent to 36 percent. On the other hand, white noncollege voters—28 percent of the total—backed Trump by 25 points, 58 percent to 33 percent.

In 2020, we project that white noncollege eligible voters should decline by around 2 points relative to 2016, and white college eligible voters should decline by 0.6 points. Black eligible voters should be stable, but Hispanics should increase by around 2 points and Asians/other race by half a point. These changes favor the Democrats.

Trump’s target in the state for 2020 would presumably be to significantly increase his 25-point margin among white noncollege voters from 2016. But even increasing this margin by 10 points would leave him 6 points behind in the state, all else remaining equal. A more effective but more difficult option would be improving his performance among Hispanics, Asians, and voters of other races by 15 margin points, which would narrow his projected deficit to 2 points. Thus, some combination of large shifts among key groups of New Mexicans is probably necessary for him to make the state competitive.

The Democratic candidate in 2020 could expand Clinton’s 2016 margin in the state by around a point by holding 2016 voting patterns constant, given underlying demographic changes in the structure of the eligible electorate. If he or she expanded Clinton’s 2016 lead by 15 points among Hispanics, Asians, and voters of other races plus boosted turnout of these groups, the projected 2020 Democratic margin would balloon into the mid-to-high teens.

Southwest summary

The five Southwest swing states are all marked by fast growth and relatively large populations of nonwhite voters, ranging from a low of 19 percent in Colorado to a high of 47 percent in New Mexico. However, nonwhite voters in these states are dominated by Hispanics, who do not vote as heavily Democratic as Black voters. The Democrats’ base of nonwhite voters in these states is supplemented by an advantage among white college-graduate populations, with the exception of Texas (though the Republicans’ advantage in Texas has been declining over time). Trump’s strongest group here, as in the rest of the country, are white noncollege voters.

In 2020, our estimates indicate that these states should see 2- to 3-point increases in the percent of nonwhite eligible voters relative to 2016, roughly balanced by 2- to 3-point declines in the percent of white noncollege eligible voters. These are all changes that favor the Democratic candidate and should help him or her keep the Democratic advantages in Colorado, Nevada, and New Mexico as well as narrow the 2016 deficit in Arizona. However, the Republicans’ advantage is too large in Texas for these changes to make a significant difference in the projected outcome for Democrats.

Democrats will seek to build on their advantage from ongoing demographic shifts by enhancing their margins among white college graduates and Hispanics and increasing the turnout levels of the latter. These improvements are most necessary for them to capture Arizona and to even make Texas competitive. And in Texas, the Democrats’ white noncollege deficit is so large that some diminution of Trump’s advantage among these voters is also probably necessary.

For Trump, he has a tall order in Colorado, Nevada, and New Mexico given Democrats’ previous advantages in these states and recent trends. There are paths for him in these states, but they are difficult. And in Arizona—key to Democrats’ electoral vote plans—he needs to safeguard his 2016 victory by enhancing his existing strong support among white noncollege voters. As for Texas, without which his electoral plans would fall apart, he needs to stem possible large-scale deterioration of support among white college graduates and Hispanics. If he can do that and keep white noncollege voters’ support, he should be fairly safe in the state.

The New South

The New South includes four states that are likely to be in play between Trump and the Democratic nominee:

• Florida: 29 electoral votes
• Georgia: 16 electoral votes
• North Carolina: 15 electoral votes
• Virginia: 13 electoral votes

Together, these four New South states have 73 electoral votes. In 2016, Trump carried all of them except Virginia. Of the three states that Trump carried, Florida is the real prize for the Democrats and would likely allow for a Democratic electoral victory. Georgia and North Carolina also represent real risks for Trump, though they were not as close as Florida in 2016. Virginia is a must-hold for the Democrats, and based on recent trends, it looks like a heavy lift for Trump.

These New South states are all fast-growing relative to the national average. They also have relatively large and growing nonwhite populations and, importantly, have large Black populations, who vote especially heavily for the Democrats. Overall, these New South states present a demographic profile and growth dynamic that is more favorable for the Democratic nominee than in the Midwest/Rust Belt swing region. That said, Trump enjoys an advantage from exceptionally conservative white populations in these states, which hobble Democrats’ efforts. We now provide a detailed discussion of these states in descending order of electoral votes.

Florida: 29 electoral votes

Trump won Florida by a single percentage point in 2016, following on Obama’s 1-point and 3-point victories in 2012 and 2008, respectively. However, Republicans won by 5 points in 2004 and by a bitterly disputed 0.01 points in 2000, which swung the election to former President George W. Bush. To say Florida is a vigorously contested state is an understatement.

In a year with several Democratic successes, the party had a relatively poor showing in Florida in 2018. Republicans won the House popular vote by more than 5 points, though Democrats did flip two GOP-held House seats. The Democrats also flipped a net of six state legislative seats from the GOP. But Republicans triumphed in the governor’s race as Democrat Andrew Gillum lost to Republican Ron DeSantis by 0.4 points. And Republican Rick Scott defeated incumbent Democratic Sen. Ben Nelson by an even smaller 0.12 points.

The closeness of the 2016 election, as well as the two statewide contests in 2018, certainly give the Democrats reasons to think they can push the needle back over to their side in 2020. But the GOP’s winning streak in the state gives the Trump campaign confidence they can hold off the charge. Adding to this confidence, Trump is currently running a negative net approval rating in the state of just -1, relatively good for Trump’s standing in 2020 swing states.

Florida has a substantial nonwhite population that made up 33 percent of voters in 2016. These voters were 17 percent Hispanic; 13 percent Black; and 3 percent Asian/other races and supported Clinton by, respectively, 78 points, 20 points, and 6 points. Florida’s white college graduates, 24 percent of voters, supported Trump by 7 points—52 percent to 45 percent—which was considerably less than Trump’s support among white noncollege voters, 43 percent of the total, who backed him by 30 points, at 63 percent to 33 percent.

We expect white noncollege eligible voters in 2020 to decline by more than 2 points relative to 2016, while white college-graduate and Black eligible voters should remain roughly stable. Hispanics should increase by more than 2 points, and Asians/other races should increase by 0.4 points. These changes in the underlying demographic structure of the electorate are enough for the Democratic candidate to knock half a point off Trump’s lead in 2020—making a very close state closer—even if voting patterns from 2016 remain in force.

Given the extreme narrowness of Trump’s victory in 2016, and the projected deterioration in his margin from demographic change, Trump needs at minimum to hold his 2016 levels of support from various demographic groups. But that’s probably not an adequate safeguard, even if attained, given possible changes in turnout patterns. The logical place for him to concentrate would be on white noncollege voters, his friendliest group. A 10-point margin shift in his favor among these voters would move his projected advantage in the state up to almost 5 points, all other voting patterns remaining the same. Another possibility is moving Hispanics—a group that includes the relatively conservative Cuban American population—as well as Asians and those of other races in his direction; a 15-point margin shift among these groups would boost his projected advantage to around 4 points.

For the Democratic candidate, a winning coalition could be assembled in several different ways. One possibility would be to move Black turnout and support back to 2012 levels, which projects to a 1-point Democratic victory, all else remaining equal. Even more effective would be a 10-point pro-Democratic margin shift among white college graduates (going from -7 to +3) that would yield a 2-point victory in the state. A 15-point pro-Democratic swing among Hispanics, Asians, and voters of other races, would also work, taking the Democrats’ projected victory margin to around 3 points. Trump will have to guard against a number of potential avenues available to the Democrats for taking back Florida, as we would expect given the closeness of the 2016 race.

Georgia: 16 electoral votes

Trump won Georgia by 5 points in 2016. This was a decline from Romney’s 8-point victory in 2012, making the trend in the state similar to that in Arizona and Texas. Democrats hope to build on this trend and make the state even closer in 2020.

Republicans managed to hold off several attempted incursions by Democrats in 2018. Republicans won the House popular vote by slightly less than 5 points, but the Democrats did flip one GOP-held House seat. The Democrats also flipped a net of 13 state legislative seats from the GOP. But Republicans won the marquee governor’s race in the state, as Republican Brian Kemp defeated Democrat Stacey Abrams by a close 1.4 points. This was the best performance by a Democrat in a Georgia governor’s race in this century, however.

These trends make the Democrats hopeful they can take the state in 2020. But the fact that the state has come no closer than 5 points in the past three elections makes the Trump campaign believe they can hold the line. Adding to this confidence, Trump is currently running a negative net approval rating in the state of -2—not great, but still better than his ratings in many other 2020 swing states.

Georgia’s large nonwhite population—38 percent of the state’s voters in 2016—is dominated by Blacks. Blacks comprised 31 percent of the voting electorate, compared with 3 percent for Hispanics and just less than 4 percent for Asians and other races. These groups supported Clinton by 76 points, 17 points, and 6 points, respectively. Meanwhile, Georgia’s white college graduates, 25 percent of voters, strongly supported Trump by 24 points, at 59 percent to 35 percent. But white noncollege voters were even stronger in their support, giving him a lopsided 63-point margin of 80 percent to 17 percent.

Our estimates indicate that white noncollege eligible voters in 2020 should decline by almost 2 points relative to 2016, while white college graduates should remain roughly stable. Both Black and Hispanic eligible voters should increase by almost a point, while Asians and other races should increase by half a point. These underlying demographic changes are enough to knock almost 2 points off the Democratic candidate’s projected disadvantage in 2020, all 2016 voting patterns remaining the same.

Given the relative closeness of Trump’s victory in 2016 plus the Democrats’ projected bonus from demographic change, Trump will seek to go beyond holding his 2016 levels of support from various demographic groups. Perhaps it’s too much to ask to increase his margin among white noncollege voters over his already mammoth 63-point advantage. But white college voters were also strong for him—and increasing his margin among them by 10 points would project to a 6-point victory in 2020.

For the Democratic candidate, the Black vote in Georgia will loom large. If the Democratic candidate could return Black turnout to 2012 levels, it would move the race to within 1.5 points of victory, all else equal. And if both Black turnout and support matched 2012 levels, it would actually produce a narrow Democratic victory. A 10-point pro-Democratic margin shift among white college graduates would be similar in effect to the increased Black turnout scenario—narrowing the gap but not quite producing victory—while shaving Trump’s immense white noncollege margin by 10 points would, in and of itself, project to a very close Democratic victory.

North Carolina: 15 electoral votes

Trump won North Carolina by just less than 4 points in 2016. This follows Romney’s narrow 2-point win over Obama in 2012 and Obama’s even narrower victory by one-third of a percentage point in 2008. All these performances were dramatically better for the Democrats compared with losing the state by 12 points in 2004 and 13 points in 2000.

Republicans continued to dominate the state in 2018, though Democrats made some progress. Republicans did relatively less well in the House popular vote, narrowly winning it by less than 2 points, but they succeeded in holding on to all GOP-held House seats. But Democrats flipped a net of 16 state legislative seats and broke Republican supermajorities in both chambers. This is of considerable significance because North Carolina’s governor is currently a Democrat.

These trends give the Democrats hope they can take the state in 2020. The Trump campaign, on the other hand, is well prepared to defend North Carolina’s 15 electoral voters—essential for their coalition—even though Trump’s current net job approval rating in the state of -3 is edging into danger territory.

North Carolina’s large nonwhite population accounted for 28 percent of voters in 2016. As in Georgia, Blacks in North Carolina dominate the nonwhite vote, representing 22 percent of all voters, compared with 3 percent for Hispanics and just less than 4 percent for Asians and other races. Blacks supported Clinton by 76 points; Hispanics by 15 points; and Asians/other races by 2 points. White college graduates in North Carolina, who represented 28 percent of voters, supported Clinton—but it was close, giving her a 4-point advantage (49 percent to 45 percent). On the other hand, white noncollege voters—43 percent of the voting electorate—gave Trump a whopping advantage of 51 points (74 percent to 23 percent).

We expect white noncollege eligible voters in 2020 to decline more than 2 points relative to 2016, while white college graduates should go up very slightly. Hispanics should increase by a point; Black eligible voters by half a point; and Asians/other races also by half a point. If 2016 voting patterns remain the same, these underlying demographic changes in the eligible electorate would be enough to reduce the Democratic candidate’s projected 2020 deficit in the state by almost 2 points.

As with Georgia, given the relative closeness of Trump’s victory in 2016 plus the projected effect of demographic change, Trump probably needs to go beyond holding his 2016 levels of group support. Increasing his margin among white college-educated voters by 10 points would yield a 5-point victory in 2020, all else equal, while increasing his already-huge lead among white noncollege voters by the same amount would project to a 6-point margin.

For the Democratic candidate, the Black vote, as in Georgia, will have great importance. If Black turnout in 2020 matches 2012 levels (there was a large decline in 2016) that would actually project to a Democratic victory of just less than a percentage point, all else equal. Matching Black support to 2012 levels would further boost the Democrats’ margin. A 10-point pro-Democratic margin shift among North Carolina’s liberalizing white college-graduate population—going from +4 to +14—would project to a narrow victory of the same magnitude as the increased Black turnout scenario. Decreasing Trump’s very large margin among white noncollege voters by 10 points would project to a larger victory.

Virginia: 13 electoral votes

Clinton won Virginia by 5 points in 2016, following on Obama’s 4- and 6-point victories in 2012 and 2008, respectively. These victories represent an impressive breakthrough for the Democrats: Prior to this streak, Republicans had carried the state in every presidential election since 1964.

Republicans did not have much to celebrate in 2018. They lost the House popular vote by an extraordinary 14 points and lost no fewer than three GOP-held House seats. Incumbent Democratic Sen. Tim Kaine also won an easy reelection victory by 26 points. Democrats’ 2018 performance came on top of their 2017 off-year state legislative wins, flipping 15 seats in the lower chamber as well as electing Democrat Ralph Northam governor by 9 points.

The Democratic candidate in 2020 looks strong given these positive trends. Trump’s path looks difficult, especially considering his current negative net job approval rating in the state of -12.

Nonwhites comprised 30 percent of Virginia voters in 2016, most of which were Black, at 19 percent, with Hispanics at 5 percent and Asians and other races at 7 percent. Blacks supported Clinton by 78 points; Hispanics by 32 points; and Asians and other races by 24 points. In addition, white college graduates, one-third of voters, backed Clinton by 8 points (51 percent to 43 percent). The bright spot for Trump was white noncollege voters, 37 percent of the voting electorate, who favored him by 41 points (68 percent to 27 percent).

Our estimates indicate that Blacks and white college eligible voters should remain stable as a share of all eligible voters in 2020, while Hispanics and Asians and other races should each increase by about a point. The sole declining group will be white noncollege voters, who are projected to decline by around 2 points. All these changes favor the Democrats.

The logical strategic choice for Trump would be to pump up his margin among the group that is far and away the most favorable to him in the state: white noncollege voters. However, even a 10-point margin shift in Trump’s direction among this demographic would still leave him 3 points behind in the state, all else remaining the same. To succeed, he would probably need to also reduce his deficit among white college voters significantly, which is a much more challenging target.

The Democratic candidate could expand Clinton’s 2016 margin by a point simply by holding Democratic margins at their 2016 levels, due to underlying demographic changes in the eligible electorate. Increasing Black turnout to 2012 levels would add a point to the Democratic candidate’s margin; furthering their lead by 15 points among Hispanics, Asians, and voters of other races would increase their lead by around two points; and expanding their already strong lead among white college graduates by 10 margin points would add 3 points to their projected advantage.

New South summary

The four New South swing states are all marked by fast growth and relatively large percentages of nonwhite voters, ranging from a low of 28 percent in North Carolina in 2016 to a high of 38 percent in Georgia. Moreover, with the important exception of Florida, nonwhite voting populations in these states are mostly Black—the most heavily Democratic-voting constituency in the country. However, in the key states of Florida and Georgia, white college-graduate voters backed Trump in 2016, making these states different from most other swing states. And white noncollege voters were very strong indeed for Trump in all these states, further hampering Democratic efforts.

Our estimates indicate that in 2020, white noncollege eligible voters in these states should decline by around 2 points relative to 2016 balanced by commensurate increases in nonwhite eligible voters. In North Carolina and especially Georgia, increases in Black eligible voters will be significant, but increases in Hispanics as well as Asians and other races dominate overall. These changes favor the Democratic candidate and should help bring him or her closer to Trump in Florida, Georgia, and North Carolina.

Democrats will seek to build on their advantage from ongoing demographic shifts by increasing Black turnout and enhancing their margins among white college graduates, Hispanics, and Asians/other races. Such improvements could plausibly bring them victory in North Carolina, Georgia, or even Florida, given the extreme closeness of the state.

Trump faces a difficult challenge in Virginia, and it is not clear how strongly his campaign will contest the state. His main task is clearly to safeguard his 2016 victories in North Carolina, Georgia, and particularly Florida, without which his electoral plans will fall apart. To do this, he will seek to increase his already gaudy numbers among white noncollege voters as well as make a run at white college voters in several states. And, at least in Florida, he will try to turn more Hispanics in his direction. It is fair to say that continued success in this region is central to his reelection effort.

And yes, there will be a quiz.