Credit...Peter van Agtmael/Magnum PhotosBy Thomas B. Edsall
Mr. Edsall contributes a weekly column from Washington, D.C., on politics, demographics and inequality.
Donald Trump’s dominance of the political stage for the past seven years galvanized what had been a slow-burning realignment, creating a profound upheaval in the electorate and in both the Democratic and Republican parties.
The support Trump received in rural communities and the animosity he provoked among well-educated suburbanites accelerated the ongoing inversion — on measures of income, education and geographic region — of white Democratic and Republican voters. (White voters make up 67 percent of the electorate.)
In 2018, according to ProximityOne, a website that analyzes the demographics of congressional districts, Democratic members of Congress represented 74 of the 100 most affluent districts, including 24 of the top 25. Conversely, Republican members of Congress represented 54 of the 100 districts with the lowest household income. The median household income in districts represented by Democrats was $66,829, which is $10,324 more than the median for districts represented by Republicans, at $56,505.
The 2018 data stands in contrast to the income pattern a half-century ago. In 1973, Republicans held 63 of the 100 highest-income districts and Democrats held 73 of the 100 lowest-income districts.
These trends prompted Nolan McCarty, a political scientist at Princeton, to comment in an email that the Democrats are mostly the party of the master’s degree — modestly advantaged economically but not exactly elite. On the flip side, the Republicans are the party of the associate degree (a two-year college degree), less educated than the Democrats but not exactly the proletariat.
Richard Pildes, a law professor at N.Y.U., argued that politics throughout the Western democracies is in recent years in the midst of the most dramatic reconfiguration of the political parties and their bases of support in seventy or so years. Since the New Deal in the United States and WWII in Western Europe, the base of the dominant parties of the left was less affluent, less highly educated voters; the dominant parties of the right drew their primary support from higher income, more highly educated voters.
Now, Pildes continued, “we are witnessing the complete inversion of that pattern, and the question is whether this is a temporary or more enduring realignment of the political parties throughout the West.”
In his email, Pildes noted that in the 1940s Democratic candidates received twenty-two points less support from voters in the top ten percent of the income bracket than from those in the bottom ninety percent. By 2012, that gap had dropped to only an eight-point difference and in 2016, voters in the top ten percent had become eight points more likely to vote for Democratic candidates. Similarly, in the 1940s, those with university degrees in the United States were twenty points less likely to vote for Democrats, while in 2000 there was no difference and by 2016, they were thirteen points more likely to vote for Democrats.
The ramifications of these developments, which predate Trump’s entry into presidential politics in 2015, “radiate throughout the electoral process in the United States,” Pildes argued:
Take the Electoral College: for most of the time from the 1950s until 2016, it was actually biased toward the Democrats. But in 2016, it suddenly became strongly biased toward the Republicans, and 2020 added even more to that bias.
At the same time, there are counter-developments more favorable to the left.
Nicholas Stephanopoulos, a law professor at Harvard who focuses on redistricting and demographic trends, argued in an email that “the country’s political geography is now less pro-Republican.” While “the conventional wisdom has it that Democrats are disadvantaged in redistricting because of their inefficient over-concentration in cities,” he continued, “the Trump era seems to have changed the country’s political geography in ways that are beneficial to Democrats.”
Trump, Stephanopoulos continued, modestly reduced the enormous Democratic edge in cities, thus undoing some of this packing of Democratic voters. Trump also did significantly better in rural areas, to the point that some of them are about as red (and so as packed with Republicans) as cities are blue. And Trump bled support in the suburbs, so that the country’s most populous and competitive areas now lean toward the Democrats instead of the Republicans.
As a result, Stephanopoulos argued, the U.S. House will likely be close to unbiased in partisan terms in 2022. A group of scholars peg the likely bias at around 3 percent pro-Republican, while Nate Silver’s model, which incorporates additional variables like incumbency and polling, thinks the likely bias will be around 1 percent pro-Democratic.
Republicans won 234 seats in 2012 despite the fact that Democrats won, by 2 percent, a majority of votes cast in House elections, according to Stephanopoulos, “but Nate Silver now thinks that Republicans will win the national House vote by 5 percent in 2022, yet only pick up the same 234 seats they got in 2012.”
Robert M. Stein, a political scientist at Rice University, agrees with Stephanopolous and cites trends in Texas to show the pro-Democratic shift:
Consider the Texas Republican Party’s redistricting plan in 2010 and its durability over the last decade. Beginning in 2010 Republicans held a 100 to 50 seat advantage in the Texas House of Representatives. By 2020, this margin had shrunk to 83-67. In each biennial election since 2010, Democrats picked up House seats, mostly in suburban and exurban areas of the state.
The shift, Stein continued, was largely driven by the changing demography of the state. Another source of this shift can be laid at the feet of candidates like Donald Trump and Ted Cruz. The result, at least in Texas, is that some of the most competitive areas (districts) in the state are not the big cities, but exurban and suburban counties including Collin, Denton, Fort Bend and Williamson. Prior to 2016 voters in these counties were trending Republican; now they are leaning Democratic or tossups.
Brian Schaffner, a political scientist at Tufts, cited surveys conducted by the Cooperative Election Study from 2010 to 2020 showing that “one of the most significant shifts we see in our data is increasing Democratic strength in suburbs, especially since the early 2010s.”
Schaffner provided data from the study showing that the Democratic share of the two-party vote rose from 54.5 to 63.5 percent in urban areas over the decade and remained low — 35.2 to 36.1 percent — in rural America. The biggest shift, 12.5 points, was in suburban areas, which went from 41.8 percent Democratic in 2010 to 54.3 percent in 2020.
Nolan McCarty suggested that these trends may prove beneficial to the Democratic Party:
The natural tilt of our single-member district system has shifted away from the Republicans as the rural vote moves toward the Republicans and the suburbs move toward the Democrats. But it is not clear what the aggregate effects of those shifts will be. It should help the House Democrats in November but it is not clear how much.
The effects of these shifts on the Senate and Electoral College, McCarty continued, will be slower in the short term but could eventually become significant: “Once such changes push states like Georgia, Texas and North Carolina sufficiently toward the Democrats, they would be the party with the structural advantage in the Electoral College and Senate.”
Jonathan Rodden, a political scientist at Stanford, noted in an email the possibility that very recent changes in suburban voting will hurt the future prospects of the Republican Party:
The most noteworthy change to political geography in 2020 was the success of Biden in pivotal suburban areas. In the most recent round of redistricting, when examining proposed districting plans — whether drawn by computer simulations or humans — the number of Democratic-leaning districts in a state was often greater if one added up the votes of Biden and Trump in 2020 than if one used past presidential results, Senate results, gubernatorial results, or some other down-ballot elections.
The geographic distribution of Biden votes, Rodden continued, “was more ‘efficient’ for the Democrats than that of other recent Democratic candidates.” But, he cautioned, what is unclear is whether this was a specific reaction to Donald Trump as a candidate in relatively educated suburbs, or a lasting trend in political geography that will outlive the Trump era. The latter is at least plausible, especially in the wake of the Dobbs decision, but it is too early to tell. Even in 2020, a non-trivial number of these suburban Biden voters split their tickets and voted for Republican House candidates.
I asked Rodden what it means for statewide elections in contested states if these trends continue. He replied:
This really depends on the numbers in each state, but in sun-belt states that are gaining educated and/or minority in-migrants, like Georgia and Arizona, we already have evidence that this was a pretty good trade for statewide Democrats, but in other states where in-migration is limited, like those in the Upper Midwest, this trade might work out better for statewide Republicans.
Along similar lines, William Frey, a demographer and a senior fellow at Brookings, emphasized in an email that “Biden won the suburbs in 2020, I believe largely due to his gains among minorities and college whites.” Even if Republicans and Trump made marginal gains among minority voters, the support of these voters for Democrats remained overwhelming.
In a 2021 Brookings paper, “Biden’s victory came from the suburbs,” Frey pointed to Georgia, where
Demographic shifts — including brisk growth in the state’s Democratic-leaning Black population, gains in Latino/Hispanic, and Asian Americans voters, and an increase in white college graduates, especially in the Atlanta metropolitan area — served to make the state competitive for Democrats this year.
In a separate 2022 paper, “Today’s suburbs are symbolic of America’s rising diversity: A 2020 census portrait,” Frey focuses on the continuing stream of minorities moving into the suburbs. From 1990 to 2020, Frey found, the percentage of Asian Americans living in suburbs grew from 53.4 to 63.1 percent, of Hispanics from 49.5 to 61.4 percent and of African Americans, from 36.6 to 54.3 percent, the largest increase.
Has geographic division, pitting a disproportionately rural Republican Party against an urban Democratic Party, added a new dimension to polarization making consensus and cooperation even more difficult?
Along similar lines, William Frey, a demographer and a senior fellow at Brookings, emphasized in an email that “Biden won the suburbs in 2020, I believe largely due to his gains among minorities and college whites.” Even if Republicans and Trump made marginal gains among minority voters, the support of these voters for Democrats remained overwhelming.
In a 2021 Brookings paper, “Biden’s victory came from the suburbs,” Frey pointed to Georgia, where
Demographic shifts — including brisk growth in the state’s Democratic-leaning Black population, gains in Latino/Hispanic, and Asian Americans voters, and an increase in white college graduates, especially in the Atlanta metropolitan area — served to make the state competitive for Democrats this year.
In a separate 2022 paper, “Today’s suburbs are symbolic of America’s rising diversity: A 2020 census portrait,” Frey focuses on the continuing stream of minorities moving into the suburbs. From 1990 to 2020, Frey found, the percentage of Asian Americans living in suburbs grew from 53.4 to 63.1 percent, of Hispanics from 49.5 to 61.4 percent and of African Americans, from 36.6 to 54.3 percent, the largest increase.
Has geographic division, pitting a disproportionately rural Republican Party against an urban Democratic Party, added a new dimension to polarization making consensus and cooperation even more difficult?
I posed a series of questions to an eclectic group of political scholars.
Frances Lee, a political scientist at Princeton, replied by email:
Rather than claiming that the G.O.P. is becoming the party of the working class, what I see is a long-term trend away from a party system organized along class lines. Knowing that a person is wealthy (or low income) isn’t very predictive of what party that person will prefer. The parties are much better sorted by other factors — region, religion, race — than by social class.
This isn’t a new phenomenon, Lee noted, but Trump intensified these divisions: “Trump’s candidacy and presidency accelerated pre-existing trends undercutting the class basis of the parties. For a Republican, Trump had unusual appeal to working-class voters and was unusually alienating to well-off suburbanites.”
James Druckman, a political scientist at Northwestern University, draws an interesting distinction: “I do think the perception in the country is that Republicans are working class but not necessarily for economic reasons directly but rather because of diffuse feelings of injustice translated into rhetoric about mistreatment, unfairness and immigrants taking jobs.”
At the same time, Druckman contended:
Democrats are vulnerable to charges of being the party of the elite for two reasons — one is that a small strain of the party is made up of extreme progressives who offer rhetoric that can be alienating when too wrapped up in politically correct language. Second, the growing anti-intellectualism in parts of the Republican Party reflects the significant degree of education polarization we observe.
Herbert Kitschelt, a political scientist at Duke, rejects some recent attempts at classification:
Are the Democrats the party of the elites? Yes and no. It is the case that high-income high-education professionals in the last 20 years have moved increasingly to the Democratic Party but these are people most of whom are on the moderate wing of the party. That is to say, they embrace a mildly redistributive agenda on economic issues such as Social Security, universal health care, and support for families with children, and a mildly libertarian social agenda on questions of abortion, family relations, gender relations and ethnic relations.
These moderate, mainstream Democrats are far removed from the more radical, progressive wing and its agenda on identity, diversity, equity, and social transformation. The real driving force of the progressive wing of the Democratic Party are occupational strata that are characterized by low- to middle-incomes and high education. These progressive voters primarily work in social and cultural services, in large urban areas.
This progressive constituency, Kitschelt argued, is quantitatively more important for the Democratic electorate than the high-education high-income more moderate segment. By embracing the agenda of “defund the police” and cultural transformation of the schools, this progressive constituency puts itself at odds with many lower- and middle-income families across all ethnic groups.
Insofar as the Democratic Party adopts the progressive agenda, Kitschelt wrote, it endangers “its electoral rainbow coalition,” noting that both African American and Hispanic families “are highly concerned about improving the police, not dismantling the police” and about “the quality of basic school instruction.”
On the Republican side, Kitschelt argues that the core element is not “working class” in any conventional sense of the phrase at all: It is low education, but relatively high-income people. These voters are overwhelmingly white, and many are of the evangelical religious conviction. In occupational terms, they are concentrated in small business, both owners and core employees, in sectors such as construction, crafts, real estate, small retail, personal services and agriculture.
Kitschelt continued: “Many of these citizens tend to live in suburban and rural areas. They are the true spearhead of Republican activism, and especially of the Trumpist persuasion.”
Pildes addressed these issues in his October 2021 paper, “Political Fragmentation in Democracies of the West.”
“The domination of the parties of the left by the more highly educated,” he wrote, “in combination with these cultural conflicts and policy differences, are an important element in the shift of the less educated, less affluent voters away from the parties of the left.”
Pildes cites American National Elections Studies data on white voters in the 2016 election showing that Trump won among all income categories of whites making less than $175,000, while Hillary Clinton won only among whites who made in excess of $175,000.
Pildes contended that defections from the Democratic Party among conservative and moderate minority voters pose a significant threat to the long-term viability of the party:
Democratic support plunged from 49 percent to 27 percent among Hispanic conservatives between 2012 and 2020 and from 69 percent to 65 percent among Hispanic moderates. These changes suggest that ideology, rather than identity, is beginning to provide more of a voting basis among some Hispanics. If a marginally greater number of working-class Latino or Black voters start to vote the way that white working-class voters do, the ability of the Democratic Party to win national elections will be severely weakened.
Bart Bonikowski, a professor of sociology and politics at N.Y.U., noted in an email that “the claim that the Republicans are becoming a party of the ‘working class’ is mistaken.” Not only are a majority of working class African Americans and Hispanics Democratic, but, “more accurately, the Republicans have become a party of disaffected white voters, many of whom hold resentments against ethnoracial minorities and a waning commitment to liberal democratic values.” Given “the built-in biases of the Electoral College and Senate — along with gerrymandering and voter disenfranchisement — states with larger shares of noncollege whites will continue to exert outsized influence on U.S. politics, persistently disadvantaging Democrats even when their candidates and policies are broadly popular.”
Robert Saldin, a political scientist at the University of Montana, argued by email that “Geographic polarization, or the urban-rural divide, is arguably the most defining feature of American politics.” Over the past 20 years, he continued, “the Democratic Party has hemorrhaged support in the countryside. They’ve got a five-alarm fire in rural America, but much of the party’s elite doesn’t even see the smoke.”
For the Democrats, in Saldin’s view, trading the countryside for the cities has come at a political cost even if the party routinely wins many more total votes than the G.O.P. nationally. That’s because geography plays an outsized role in our political system, particularly in the Electoral College and the Senate.
Consider the Dakotas, Saldin wrote:
It wasn’t that long ago that their congressional delegations were packed with Democrats, but that’s inconceivable now. And to the extent that the same thing is happening in other low-population states, this presents a real problem for Democrats in the Senate.
Saldin suggested:
Here’s another way of conceptualizing it. Idaho, Montana, North Dakota, South Dakota and Wyoming have less than 2 percent of the national population, but their ten senators have the same collective power in the Senate as those representing the five most populous states, California, Texas, Florida, New York and Pennsylvania. If a party managed a clean sweep in those five big-box states in flyover country, that would comprise 20 percent of what you need for a Senate majority before you even look at the other 98 percent of the country. The G.O.P. is now very close to accomplishing that feat, with Montana’s Jon Tester the last Senate Democrat standing in those states.
Barring an extraordinary economic turnaround or still more explosive disclosures of criminal malfeasance by Trump, these demographic trends may have a modest effect on the outcome on Election Day in November. They do, however, suggest that the balance of political power is more fluid than widely recognized. It should undermine the confidence of those predicting victory for either the left or the right in 2024.
Frances Lee, a political scientist at Princeton, replied by email:
Rather than claiming that the G.O.P. is becoming the party of the working class, what I see is a long-term trend away from a party system organized along class lines. Knowing that a person is wealthy (or low income) isn’t very predictive of what party that person will prefer. The parties are much better sorted by other factors — region, religion, race — than by social class.
This isn’t a new phenomenon, Lee noted, but Trump intensified these divisions: “Trump’s candidacy and presidency accelerated pre-existing trends undercutting the class basis of the parties. For a Republican, Trump had unusual appeal to working-class voters and was unusually alienating to well-off suburbanites.”
James Druckman, a political scientist at Northwestern University, draws an interesting distinction: “I do think the perception in the country is that Republicans are working class but not necessarily for economic reasons directly but rather because of diffuse feelings of injustice translated into rhetoric about mistreatment, unfairness and immigrants taking jobs.”
At the same time, Druckman contended:
Democrats are vulnerable to charges of being the party of the elite for two reasons — one is that a small strain of the party is made up of extreme progressives who offer rhetoric that can be alienating when too wrapped up in politically correct language. Second, the growing anti-intellectualism in parts of the Republican Party reflects the significant degree of education polarization we observe.
Herbert Kitschelt, a political scientist at Duke, rejects some recent attempts at classification:
Are the Democrats the party of the elites? Yes and no. It is the case that high-income high-education professionals in the last 20 years have moved increasingly to the Democratic Party but these are people most of whom are on the moderate wing of the party. That is to say, they embrace a mildly redistributive agenda on economic issues such as Social Security, universal health care, and support for families with children, and a mildly libertarian social agenda on questions of abortion, family relations, gender relations and ethnic relations.
These moderate, mainstream Democrats are far removed from the more radical, progressive wing and its agenda on identity, diversity, equity, and social transformation. The real driving force of the progressive wing of the Democratic Party are occupational strata that are characterized by low- to middle-incomes and high education. These progressive voters primarily work in social and cultural services, in large urban areas.
This progressive constituency, Kitschelt argued, is quantitatively more important for the Democratic electorate than the high-education high-income more moderate segment. By embracing the agenda of “defund the police” and cultural transformation of the schools, this progressive constituency puts itself at odds with many lower- and middle-income families across all ethnic groups.
Insofar as the Democratic Party adopts the progressive agenda, Kitschelt wrote, it endangers “its electoral rainbow coalition,” noting that both African American and Hispanic families “are highly concerned about improving the police, not dismantling the police” and about “the quality of basic school instruction.”
On the Republican side, Kitschelt argues that the core element is not “working class” in any conventional sense of the phrase at all: It is low education, but relatively high-income people. These voters are overwhelmingly white, and many are of the evangelical religious conviction. In occupational terms, they are concentrated in small business, both owners and core employees, in sectors such as construction, crafts, real estate, small retail, personal services and agriculture.
Kitschelt continued: “Many of these citizens tend to live in suburban and rural areas. They are the true spearhead of Republican activism, and especially of the Trumpist persuasion.”
Pildes addressed these issues in his October 2021 paper, “Political Fragmentation in Democracies of the West.”
“The domination of the parties of the left by the more highly educated,” he wrote, “in combination with these cultural conflicts and policy differences, are an important element in the shift of the less educated, less affluent voters away from the parties of the left.”
Pildes cites American National Elections Studies data on white voters in the 2016 election showing that Trump won among all income categories of whites making less than $175,000, while Hillary Clinton won only among whites who made in excess of $175,000.
Pildes contended that defections from the Democratic Party among conservative and moderate minority voters pose a significant threat to the long-term viability of the party:
Democratic support plunged from 49 percent to 27 percent among Hispanic conservatives between 2012 and 2020 and from 69 percent to 65 percent among Hispanic moderates. These changes suggest that ideology, rather than identity, is beginning to provide more of a voting basis among some Hispanics. If a marginally greater number of working-class Latino or Black voters start to vote the way that white working-class voters do, the ability of the Democratic Party to win national elections will be severely weakened.
Bart Bonikowski, a professor of sociology and politics at N.Y.U., noted in an email that “the claim that the Republicans are becoming a party of the ‘working class’ is mistaken.” Not only are a majority of working class African Americans and Hispanics Democratic, but, “more accurately, the Republicans have become a party of disaffected white voters, many of whom hold resentments against ethnoracial minorities and a waning commitment to liberal democratic values.” Given “the built-in biases of the Electoral College and Senate — along with gerrymandering and voter disenfranchisement — states with larger shares of noncollege whites will continue to exert outsized influence on U.S. politics, persistently disadvantaging Democrats even when their candidates and policies are broadly popular.”
Robert Saldin, a political scientist at the University of Montana, argued by email that “Geographic polarization, or the urban-rural divide, is arguably the most defining feature of American politics.” Over the past 20 years, he continued, “the Democratic Party has hemorrhaged support in the countryside. They’ve got a five-alarm fire in rural America, but much of the party’s elite doesn’t even see the smoke.”
For the Democrats, in Saldin’s view, trading the countryside for the cities has come at a political cost even if the party routinely wins many more total votes than the G.O.P. nationally. That’s because geography plays an outsized role in our political system, particularly in the Electoral College and the Senate.
Consider the Dakotas, Saldin wrote:
It wasn’t that long ago that their congressional delegations were packed with Democrats, but that’s inconceivable now. And to the extent that the same thing is happening in other low-population states, this presents a real problem for Democrats in the Senate.
Saldin suggested:
Here’s another way of conceptualizing it. Idaho, Montana, North Dakota, South Dakota and Wyoming have less than 2 percent of the national population, but their ten senators have the same collective power in the Senate as those representing the five most populous states, California, Texas, Florida, New York and Pennsylvania. If a party managed a clean sweep in those five big-box states in flyover country, that would comprise 20 percent of what you need for a Senate majority before you even look at the other 98 percent of the country. The G.O.P. is now very close to accomplishing that feat, with Montana’s Jon Tester the last Senate Democrat standing in those states.
Barring an extraordinary economic turnaround or still more explosive disclosures of criminal malfeasance by Trump, these demographic trends may have a modest effect on the outcome on Election Day in November. They do, however, suggest that the balance of political power is more fluid than widely recognized. It should undermine the confidence of those predicting victory for either the left or the right in 2024.