There are many candidate explanations for this: the relatively less important role of religion, the relatively higher level of education, the fact that living in crowded areas leads people to appreciate interdependence, etc.

However, I am looking for more serious scientific studies that seriously test these explanations, instead of just telling stories. If you know any sources, please let me know.

  • Ethnicity will come into it at some level. Urban centers tend to be considerably less white than the suburban and rural areas that encircle them. Dec 28, 2020 at 1:31
  • @Crazymoomin Urban whites are much more liberal the rural whites.
    – ohwilleke
    Dec 28, 2020 at 7:35
  • @ohwilleke It's not necessarily a conservative/liberal thing, ethnic minorities are more likely to vote democrat regardless of their actual political views. The litmus test would be to find an overwhelmingly white urban center for comparison. Dec 28, 2020 at 11:51
  • @Crazymoomin Ethnic minorities (e.g. Muslims) who are Democrats are often more conservative than white Democrats, but all Democrats are much more liberal than Republicans. This wasn't always the case but the parties are now firmly divided into liberal and conservative camps to a purity not seen at any time in recent history.
    – ohwilleke
    Dec 28, 2020 at 21:39
  • 1
    @dsollen O.K., I eventually decided to address the points raised in your comment and some others with a much extended narrative.
    – ohwilleke
    Dec 28, 2020 at 23:43

2 Answers 2


There are many candidate explanations for this: the relatively less important role of religion, the relatively higher level of education, the fact that living in crowded areas leads people to appreciate interdependence, etc. . . . I am looking for more serious scientific studies that seriously test these explanations, instead of just telling stories.

The question is basically looking for a structural functionalist explanation of partisan leanings associated with population density. And, this kind of explanation can explain some of what we see. But the political ideologies of people don't have only these kinds of structural functionalist roots.

The full story can only fully be described in a path dependent narrative in which history and inertia overcome the general considerations that apply ceteris paribus (all other things being equal). Particularly when it comes to explaining why urban ethnic and religious minorities tend to be Democratic party leaning even when their economic circumstances would suggest otherwise, it is necessary to tell stories (which is what history does), and these explanations are not less serious simply because they are historical narratives which need to be understood rather than reductive formulas and general principles.

Population Density and Partisan Lean Are Strongly But Not Perfectly Related

The place to begin is to recognize that this correlation is real and well established. Population density is a first order strongly correlated predictor of partisan preference, all other things being equal. It isn't a perfect correlation, however. A scatter plot of states by population density v. partisan lean in the 2016 election illustrates the point:

enter image description here

This doesn't just hold true in Presidential elections:

The median Republican Congressional district now has a population density 11 times smaller than the median Democratic district[.]

enter image description here

From here. Chart from here, which goes on to explain:

As shown in the chart above (in Log scale), there was a relatively strong positive correlation between density of congressional districts and the vote share of the Democratic candidate in the 2010 elections. Of densest quartile of districts with a race between a Democrat and a Republican — 105 of them, with a density of 1,935 people per square miles or more — the Democratic candidate won 89. Of the quartile of districts with the lowest densities — 98 people per square mile and below — Democratic candidates only won 23 races.

Economic productivity and population density are related

To a significant extent, the relationship between population density and partisan lean echos another well established economic relationship, which is that "productivity increases by 7% to 15% when the municipality population density doubles" that has been replicated across many cultures, economy types and time periods. Others studies suggest the same relationship but suggests a slightly weaker but still consistent relationship: "doubling density increases productivity by an average of two to four percent." See also, e.g., here.

enter image description here

Image from here.

enter image description here

52% of total GDP [is] created in America’s 20 top metropolitan areas.

The quotation and image is from here.

Economic Prosperity and Partisan Leanings Are Related

Well done multi-cultural evaluations of world values, such as the World Values Survey show that, in general, liberal and secular leaning political views thrive in circumstances that are more economically prosperous and secure, while conservative and religious political views thrive in circumstances of economic scarcity and insecurity. As David Brooks explains this works in the context of domestic politics as well:

People need a secure order to feel safe. Deprived of that, people legitimately feel cynicism and distrust, alienation and anomie. This precarity has created, in nation after nation, intense populist backlashes against the highly educated folks who have migrated to the cities and accrued significant economic, cultural and political power. Will Wilkinson of the Niskanen Center calls this the “Density Divide."

Wilkinson's analysis at the link above explains that:

[U]rbanization is a relentless, glacial social force that transforms entire societies and, in the process, generates cultural and political polarization by segregating populations along the lines of the traits that make individuals more or less responsive to the incentives that draw people to the city. I explore three such traits — ethnicity, ideology-correlated aspects of personality, and level of educational achievement — and their intricate web of relationships. The upshot is that, over the course of millions of moves over many decades, high density areas have become economically thriving multicultural havens while whiter, lower density places are facing stagnation and decline as their populations have become increasingly uniform in terms of socially conservative personality, aversion to diversity, and lower levels of education. This self-segregation of the population, I argue, created the polarized economic and cultural conditions that led to populist backlash.

Certainly, one important factor mediating the connection between partisan leaning and population density is economic prosperity and productivity, even though it isn't a perfect predictor.

In the U.S., blue counties average about double the per capita GDP of red counties. For example, even though Biden won about 53% of the popular vote in 2020, counties he won had about 70% of the U.S. GDP while counties that Trump won had about 30% of U.S. GDP (a GDP per capita ratio of 2.07).

Migration and Diversity

Migration is the primary mechanism by which economic prosperity which established regional norms and ideological world views and political understandings transmits itself to less economically prosperous people in the same area who move there to participate in the prosperity that they have not yet themselves attained. Migrants adopt local worldviews and ideologies because the whole point of them moving to these prosperous areas was to integrate themselves into the local cultural in order to ultimately share in that prosperity.

It is a similar phenomena to poor conservatives supporting abolishing estate taxes that they will never pay because they see themselves as part of that dream that they might someday face that problem, even if that is not a realistic assumption.

It is certainly true that urban areas are more racially and ethnically diverse than rural areas, on average, in the U.S., although it is a bit more complicated than that, because the reasons that areas outside the South are diverse are primarily due to voluntary acts, while the racial diversity of the South is primary rooted in the inherently involuntary institution of slavery which prevailed until 1865 there, and had de facto vestiges that continues much longer and still remain to this day in some respects.

[In the U.S. House of Representatives] Districts represented by Democrats are, together, only slightly more than half white and nearly one quarter Hispanic. Republican districts are roughly three-quarters white, and only one in nine residents is Hispanic.

From here.

The case of the American South and its impact on the larger political system and process of ideology formation in ethnic and religious minorities in the U.S. is something of a case of American exceptionalism driven by its history of slavery and racism. Basically, the more conservative party (the Democrats in the post-Civil War era and the Republicans in the post-Civil Rights movement era) interposed itself to demonize ethnic minorities as a way to bolster poor whites.

Outside the South, the ethnic diversity of urban areas is largely a function of migration driven by urban prosperity.

Immigrants to the U.S. from abroad disproportionately migrated to these region in the hopes that they should share in the economic prosperity of those regions, particularly as urban areas gained more of an economic edge relative to rural areas over time.

Outside the South, the lion's share of African-Americans migrated within the U.S. to regions outside the South for similar reasons, especially during the "Great Migration" from around 1910 to 1970 in the United States, when African Americans migrated in large numbers to industrial cities in the North for manufacturing jobs. Prior to that, African Americans were overwhelmingly residents of the American South (i.e. the former Confederate States).

These groups and a lot of other minority ethnicities, however, defied the usual trend of affiliating with the more conservative party and developing those ideologies despite economically struggling once they arrived.

In contrast, ethnic and religious minorities have avoided rural America which has been in more or less continuous economic decline (collectively, outside newly settled frontiers where there was growth because there was no agriculture or modern economic activity of note there in the eyes of the settlers) since the 1790s, since economic opportunity for newcomers has been lacking there and since they weren't welcome and that lack of welcome wasn't worth overcoming in the face of stagnant economic prospects for newcomers.

The American South, however, is another story, that spilled out to influence political ideology at a mass grass roots level nationwide among ethnic and religious minorities in the U.S. There, the longstanding legacy of slavery and its recent dismantlement (compared, for example, to the demise of serfdom in Europe) helps explain why the usual considerations have been overcome there and spilled out to other parts of the U.S.

African American alienation from the modern Republican political party can be traced, fairly clearly, to the outright racism and hostility present in the Republican party which became the new home of segregationist Southerners (the former "Dixiecrats") who wanted to keep first de jure and then de facto and private sector discrimination against African Americans in place. Republicans have, at every step, opposed the integration of public institutions, viable employment discrimination laws, and viable remedies for victims of abuse in the criminal justice system (particularly those like the sentencing disparities between powder and crack cocaine that disproportionately impact African Americans for no principled reason). Republicans have opposed African American judicial and cabinet nominees (with the notable exception of African American conservative U.S. Supreme Court Justice Thomas). Republicans have favored regressive taxation schemes and weak social welfare safety nets in ways that have done disproportionate economic harm to African Americans. Republicans have opposed history instruction in schools that highlights African American contributions. Republicans opposed making Martin Luther King, Jr. day a holiday.

Republicans have also not historically been very welcoming other conservative leaning immigrant groups like Muslims and Protestant Hispanics.

The Democratic party, in contrast (during the time period that it was the more liberal party), has had the back of African Americans on issues with strong racial overtones for the last half century or more, has consistently advanced African American economic interests, and has embraced African American advancement in partisan political party offices, as candidates for elective office, and as political appointees in executive branch and judicial branch offices. This embrace of diversity was a natural development for a party rooted in urban areas that saw surges of immigration and internal migration turning their communities into highly diverse ones united by a desire to participate in urban prosperity.

Conservative policies threatening the very existence of ethnic and racial minorities were more salient than the more general conservative leaning of these populations in the U.S., even though both poor, rural whites and poor urban non-whites faced similar economic challenges (and continue to face those challenges).

Membership In The Democratic Party Coalition Changed African Americans And Other Minorities Ideologically

This coalition, in turn, has mobilized elites in the African American community to back a Democratic party political agenda, and to advocate for and cause the realization of African American community values in the United States, such as efforts to find common cause with Hispanic immigrants, rather than seeing them as threats, and seeing environmental protection as something that personally advances their health rather than as a threat to their jobs, that would not have been natural responses in a vacuum outside the Democratic Party coalition.

Similarly, the larger Democratic Party coalition has influenced African Americans in responding to criminal violence in their communities with calls for gun control, rather than easier access to gun ownership (e.g. through restoration of the civil rights of those with criminal convictions).

Put another way, African Americans have largely repudiated the culture of the American South within which their own ethnic subculture arose, and realigned their own subculture with that of the economically successful parts of the rest of the United States. Following that switch of cultural allegiances, their own conditions of personal safety and economic security have been relegated to second order within community differences in values that also shift the entire community's values modestly in their direction, rather than reinforcing the similar although less intense values of working class white Southerners in a similar conservative direction.

Public Policy Also Changed African American Values

Public policy itself has changed African American political values.

Before formal policies at the governmental level made this so, quasi-governmental political machines provide mutual support and with that economic security to new immigrants outside the structure of formal government when the concept of a welfare state was not a mature policy concept, putting the beneficiaries of this support at ease with providing political support to this movement.

In the Great Depression, the massive Democratic party led expansion of the social safety net under FDR muted the far right movements like Fascism that grew more powerful in place that did not adopt them by making people more economically secure.

On the issue of health care, there was been a sustained period of time from the adoption of the Medicaid program, until the implementation of Obamacare, where as a matter of policy driven by the political coalitions backing these programs, those in poverty (who were more often urban minorities) had more secure access to medical care, than working class individuals with no health insurance (who were more often rural whites), a substantial share of all working class people.

Affirmative action arguably made it easier for an African Americans who are ready to go to college to obtain a higher education than an academically comparable working class white person for a sustained period of time, again building loyalty to liberal ideals by providing greater economic prospects for those won over to them.

There have been time periods in many places where African Americans have more heavily utilized the social welfare safety net available than comparable poor whites, in part due to better community knowledge of how the programs work, and in part due to a lower degree of distrust of those programs.

These policies and others like them make it possible to conclude that public policy has done more to alleviate insecurity for blacks than for whites for a sustained period of time, muting somewhat differences in absolute physical and economic well being, although this argument alone is not enough to explain the reality in the absence of the rejection of ethnic and religious minorities by the more conservative party in the U.S. two party system.

The Economic Divide Between Urban and Rural Areas In The U.S. Has Grown Wider Over The Last Couple of Generations

In the U.S., it has also been observed that deeper partisan political divisions between urban and rural areas have tracked a widening prosperity divide between urban and rural areas. For example, this account at Politico notes that:

[T]he economic divide between rural and urban America has widened. Small-business growth has slowed in rural communities since the Great Recession, and it has only worsened with Covid-19. . . . The source of our wealth is in the things we grow. But today, those things get shipped off into a vast global supply chain, where profits are siphoned off and little remains for us to save or invest. Farmers’ share of every retail food dollar has fallen from about 50 percent in 1952 to 15 percent today.

Prosperity Also Drives The Importance Of Religion And The Importance Of Religion Influences Partisan Leanings

It is hard to disentangle the relevant factors, however, because they aren't independent from each other.

For example, the extent to which people are religious also influences their partisan leanings. All other things being equal, more religious people are more conservative and more secular people are more liberal. But, as noted above, the importance of religion in the lives of people is strongly related to economic prosperity, something that has been shown both in an international context:

enter image description here

The deviations from the overall average curve tend to have historical roots because values and partisan ideas change gradually on the scale of generations, not just election to election.

The countries below the curve tend to be either currently, or recently Communist countries, where atheism was a matter of official policy, or Asian countries where Islam is rare in most of the country and the role of religion in politics in these countries is different than in the West.

Countries where there was once an established religion also tend to be more secular.

Countries with a significant Muslim population and countries with religious diversity, in contrast, tend to be above the curve.

Similar strong general trends that are nonetheless not perfect are also true domestically, as one can see from the data below:

% of adults who say religion is very important in their lives

  1. Alabama 77%
  2. Mississippi 74%
  3. Tennessee 71%
  4. Louisiana 71%
  5. Arkansas 70%
  6. South Carolina 69%
  7. West Virginia 64%
  8. Georgia 64%
  9. Oklahoma 64%
  10. Texas 63%
  11. Kentucky 63%
  12. North Carolina 62%
  13. Virginia 60%
  14. New Mexico 59%
  15. Utah 58%
  16. South Dakota 57%
  17. Missouri 56%
  18. Ohio 56%
  19. Nebraska 54%
  20. Iowa 53%
  21. Florida 53%
  22. Indiana 53%
  23. North Dakota 53%
  24. Arizona 51%
  25. Pennsylvania 51%
  26. Idaho 51%
  27. Kansas 50%
  28. New Jersey 50%
  29. Maryland 50%
  30. District of Columbia 50%
  31. Michigan 50%
  32. Illinois 50%
  33. Wyoming 49%
  34. Rhode Island 48%
  35. California 47%
  36. Colorado 47%
  37. Delaware 46%
  38. Minnesota 46%
  39. Oregon 45%
  40. New York 45%
  41. Nevada 44%
  42. Montana 44%
  43. Hawaii 44%
  44. Washington 44%
  45. Wisconsin 44%
  46. Connecticut 42%
  47. Alaska 41%
  48. Maine 34%
  49. Massachusetts 33%
  50. New Hampshire 33%
  51. Vermont 32%

with per capita GDP:

U.S. states and territories by GDP per capita (current dollars)

National rank State or territory 2018

51 Mississippi 37,948

50 Arkansas 42,454

49 West Virginia 43,053

48 Idaho 43,430

47 Alabama 45,219

46 South Carolina 45,280

45 Montana 46,609

44 Kentucky 46,898

43 New Mexico 46,954

42 Maine 47,969

41 Arizona 48,055

40 Florida 48,318

39 Oklahoma 50,613

38 Missouri 51,699

37 Michigan 53,209

36 Vermont 53,523

35 Louisiana 53,589

34 Tennessee 53,933

33 North Carolina 54,441

32 Indiana 55,172

31 Nevada 55,269

30 Utah 55,550

29 Georgia 55,832

28 Kansas 56,334

27 Oregon 56,956

26 Ohio 57,492

25 Wisconsin 57,720

24 Rhode Island 57,852

23 South Dakota 58,624

22 Iowa 59,977

21 Texas 61,167

20 Pennsylvania 61,594

— United States 62,390

19 Virginia 62,563

18 New Hampshire 63,067

17 Colorado 63,882

16 Nebraska 63,942

15 Hawaii 64,096

14 Minnesota 64,675

13 Illinois 67,268

12 Maryland 68,573

11 New Jersey 69,378

10 Wyoming 69,900

9 North Dakota 72,597

8 Alaska 73,205

7 Washington 74,182

6 California 74,205

5 Connecticut 76,342

4 Delaware 77,253

3 Massachusetts 82,480

2 New York 85,746

1 District of Columbia 200,277

  • I wonder how that GDP/religion graph looks if you use median GDP instead of average. Perhaps the US has a surprisingly high number of very rich people? Dec 28, 2020 at 11:49
  • 1
    @user253751 It wouldn't change much. The US would still have a comparable wealth per capita to countries like Japan, Australia or Western Europe, all of which are considerably less religious.
    – quarague
    Dec 28, 2020 at 15:25
  • 2
    (actually "median GDP per capita" doesn't make sense, but "median wage" does) Dec 28, 2020 at 15:28
  • 1
    The problem here is that you are trying to make sense of things on a one-dimensional scale, when US politics (and probably politics anywhere) just doesn't work like that. Just for one instance of many, under conservative/Republican, you have the religious right social conservatives (who want government to make everyone live according to their beliefs), and the libertarian "keep the damned government out of my life" ones, who are united only by their shared detestation of liberals.
    – jamesqf
    Dec 28, 2020 at 19:00
  • @jamesqf There is pretty solid empirical evidence to show that U.S. politics, while not perfectly matched to a one-dimensional scale is very well approximated by one at both the elected official level and in the rank and file. This is beyond the scope of a single short comment, however. For elected officials see: swarthmore.edu/SocSci/rvalell1/documents/Nominate_000.pdf
    – ohwilleke
    Dec 28, 2020 at 21:57

There is the big issue of correlation versus causation.

Things like higher education, higher population density and less important role of religion all have positive correlation with leaning democrat. This is scientifically rigorous. You can quantify this things (maybe not religion but the others) and then simply compute the correlation.

Causal relationships on the other hand are almost impossible to establish. On something as complicated as politics humans just aren't monocausal.

  • I agree. At a minimum, though, people should have ran a regression with these competing variables and see which ones have higher marginal explanatory power?
    – J Li
    Dec 25, 2020 at 8:52
  • @JLi Sure but then all the details start to matter. Do you measure at state or county level? Which variables do you include in your analysis? How do you put education into a numerical variable? Is leaning democrat the result of the last election, previous ones or some polling result using self identification?
    – quarague
    Dec 25, 2020 at 9:20
  • I would be happy to learn about results at all levels. If you know any I would love to learn.
    – J Li
    Dec 25, 2020 at 9:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .