2012-16 Polarization Between Metro & Rural Counties (MAP)

[Adam Griffin]

I've been wanting to look at the difference in presidential performance between rural and urban areas for some time now. I finally decided to take a look at both 2012 and 2016, with simplicity basically resulting in me examining "rural" and "not rural" areas (henceforth referred to as "metro areas"). All of the data displayed below can be found here.

Updated: the (new) maps below use the UCC delineations published by jimrtex to create 100 jurisdictions (excluding AK for now, and RI/NJ due to no rural counties; so in reality, 96) based on a county's classification.



With these boundaries established, I then began to examine each state's two jurisdictions, mapping out the difference between each group's margins in both the 2012 and 2016 presidential elections. You can find the exact margins for each state and grouping in the above spreadsheet. Below, a map averaging the difference in margin between each state's rural and metro areas for 2012/2016 can be found:



From that, I've mapped out the margins for each jurisdiction, creating a patchwork political quilt of sorts showing how each of the two groups in each state voted.

I have also attempted to apply electoral votes for these "100 states" in some sense: I assigned 1 EV to every jurisdiction and then (rather lazily but accurately given how presidential elections work) assigned the rest based on 2010 population to generate a hypothetical Electoral College outcome.

The first set of maps essentially shows the same as the polarization maps above: an average of the margin for the past 2 presidential elections.






And the same figures for 2016 alone:







And 2012's figures:







And swing:

[Adam Griffin]

Here are animated GIFs showing the margins/winner for both 2012 and 2016 in each jurisdiction:

[Adam Griffin]

Obama-Trump Districts:
NH-Metro
MO-Metro
SC-Rural
ME-Rural
WI-Rural

Romney Clinton Districts:
TX-Metro


And just in case anybody was wondering: the reason why the average EC margin is smaller than both the 2012 and 2016 EC margins is because Obama barely won MO-Metro, NH-Metro, WI-Rural and SC-Rural in 2012, while Clinton barely won TX-Metro in 2016. In terms of average, all but two of these jurisdictions (NH-Metro & MO-Metro) ended up being won by the GOP by much more than the Dem won it by in the other election, thereby awarding their EVs to the GOP on the averaged map.

[jamestroll]

Obama-Trump Districts:
ME-Rural
WI-Rural
MO-Metro

Romney Clinton Districts:
TX-Metro
AZ-Metro


And just in case anybody was wondering: the reason why the average EC margin is smaller than both the 2012 and 2016 EC margins is because Obama barely won MO-Metro in 2012, while Clinton barely won AZ-Metro & TX-Metro in 2016. In terms of average, all 3 of these jurisdictions ended up being won by the GOP by much more than the Dem won it by in the other election, thereby awarding their EVs to the GOP on the averaged map.

I am glad I took my biased lens off.. Missouri is going to be as Republican as Utah eventually!

[Adam Griffin]

Obama-Trump Districts:
ME-Rural
WI-Rural
MO-Metro

Romney Clinton Districts:
TX-Metro
AZ-Metro


And just in case anybody was wondering: the reason why the average EC margin is smaller than both the 2012 and 2016 EC margins is because Obama barely won MO-Metro in 2012, while Clinton barely won AZ-Metro & TX-Metro in 2016. In terms of average, all 3 of these jurisdictions ended up being won by the GOP by much more than the Dem won it by in the other election, thereby awarding their EVs to the GOP on the averaged map.

I am glad I took my biased lens off.. Missouri is going to be as Republican as Utah eventually!

Fun fact: MO-Rural had the largest margin in favor of Trump of any other state (Trump +54), with TN & KY (Trump +53), NE & OK (Trump +52) & WV (Trump +51) nipping at its toes!

[jamestroll]

Obama-Trump Districts:
ME-Rural
WI-Rural
MO-Metro

Romney Clinton Districts:
TX-Metro
AZ-Metro


And just in case anybody was wondering: the reason why the average EC margin is smaller than both the 2012 and 2016 EC margins is because Obama barely won MO-Metro in 2012, while Clinton barely won AZ-Metro & TX-Metro in 2016. In terms of average, all 3 of these jurisdictions ended up being won by the GOP by much more than the Dem won it by in the other election, thereby awarding their EVs to the GOP on the averaged map.

I am glad I took my biased lens off.. Missouri is going to be as Republican as Utah eventually!

Fun fact: MO-Rural had the largest margin in favor of Trump of any other state (Trump +54), with TN & KY (Trump +53), NE & OK (Trump +52) & WV (Trump +51) nipping at its toes!

haha yes! And last month in HD 144 there was literally a perfect Democratic candidate and a flawed Republican candidate in historically Democratic territory but the Democrat still lost by 5 points even if we managed to pick up an exurban seat.

If Missouri Democrats could get Platte and Clay County as solid democratic locally and Jefferson and St Charles as swing counties locally they may have a chance to compete state wide. But other wise its dead. Even if there maybe a last hurrah for constitutional officers in 2018/2020.

My feeling is that even if the state is elastic the suburbs are just too white and the rural areas are too southern cultured and too white to even be on the fringe of competitiveness.  It is not like downstate Illinois, rural Kentucky, rural West Virginia and all these other places where Trump did well yet local Democrats still held their own despite coattails. In Missouri it was a blood bath all the way down the ballot.

[America's Sweetheart ❤/𝕿𝖍𝖊 𝕭𝖔𝖔𝖙𝖞 𝖂𝖆𝖗𝖗𝖎𝖔𝖗]

Obama-Trump Districts:
ME-Rural
WI-Rural
MO-Metro

Romney Clinton Districts:
TX-Metro
AZ-Metro


And just in case anybody was wondering: the reason why the average EC margin is smaller than both the 2012 and 2016 EC margins is because Obama barely won MO-Metro in 2012, while Clinton barely won AZ-Metro & TX-Metro in 2016. In terms of average, all 3 of these jurisdictions ended up being won by the GOP by much more than the Dem won it by in the other election, thereby awarding their EVs to the GOP on the averaged map.

I am glad I took my biased lens off.. Missouri is going to be as Republican as Utah eventually!

Unless St. Louis dramatically loses population sometime in the future, I don't see how that's possible.

[muon2]

I like the analysis. I'm curious about the election of counties in IL. It looks like Sangamon is considered rural while a number of distinctly more rural counties are marked as urban (Adams, Knox, for example). Tazewell is also mostly urban population. Jimrtex's thread for UCCs has the urban fraction in each of the counties he uses.

[Adam Griffin]

I like the analysis. I'm curious about the election of counties in IL. It looks like Sangamon is considered rural while a number of distinctly more rural counties are marked as urban (Adams, Knox, for example). Tazewell is also mostly urban population. Jimrtex's thread for UCCs has the urban fraction in each of the counties he uses.

Good catch! I went back and looked at the numbers and it appears that Sangamon was included in the data as metro, but I missed it when making the maps (hand-colored, of course). There very well may be other mistakes as far as which category they've been placed in on the map/in the data, so I'm grateful to anybody who can point them out to me.


But this does bring up an important point: there are many counties where I felt quite conflicted about which category in which to include them. As I mentioned, I used a rather subjective approach in looking at these counties (with vote totals & square mileage playing big roles, along with proximity to urban clusters, relative to the state's population as a whole). A county in PA that was considered rural might be considered metro in, say, ND.

I can definitely say that one example that has been sticking out to me is St Tammany, LA (currently classified as rural). Other examples of states I'd like to revisit include PA & NC, where I feel more area could be considered metro rather than rural. It appears I have also misclassified Grenada, MS.

[muon2]

I like the analysis. I'm curious about the election of counties in IL. It looks like Sangamon is considered rural while a number of distinctly more rural counties are marked as urban (Adams, Knox, for example). Tazewell is also mostly urban population. Jimrtex's thread for UCCs has the urban fraction in each of the counties he uses.

Good catch! I went back and looked at the numbers and it appears that Sangamon was included in the data as metro, but I missed it when making the maps (hand-colored, of course). There very well may be other mistakes as far as which category they've been placed in on the map/in the data, so I'm grateful to anybody who can point them out to me.


But this does bring up an important point: there are many counties where I felt quite conflicted about which category in which to include them. As I mentioned, I used a rather subjective approach in looking at these counties (with vote totals & square mileage playing big roles, along with proximity to urban clusters). A county in PA that was considered rural might be considered metro in, say, ND.

I can definitely say that one example that has been sticking out to me is St Tammany, LA (currently classified as rural). Other examples of states I'd like to revisit include PA & NC, where I feel more area could be considered metro rather than rural. It appears I have also misclassified Grenada, MS.

I thought the UCC list would be quite helpful here. It's based on specific criteria measuring the urban vs rural population in the counties, and the urban fraction is listed for each county in the thread.

[Adam Griffin]

I like the analysis. I'm curious about the election of counties in IL. It looks like Sangamon is considered rural while a number of distinctly more rural counties are marked as urban (Adams, Knox, for example). Tazewell is also mostly urban population. Jimrtex's thread for UCCs has the urban fraction in each of the counties he uses.

Good catch! I went back and looked at the numbers and it appears that Sangamon was included in the data as metro, but I missed it when making the maps (hand-colored, of course). There very well may be other mistakes as far as which category they've been placed in on the map/in the data, so I'm grateful to anybody who can point them out to me.


But this does bring up an important point: there are many counties where I felt quite conflicted about which category in which to include them. As I mentioned, I used a rather subjective approach in looking at these counties (with vote totals & square mileage playing big roles, along with proximity to urban clusters). A county in PA that was considered rural might be considered metro in, say, ND.

I can definitely say that one example that has been sticking out to me is St Tammany, LA (currently classified as rural). Other examples of states I'd like to revisit include PA & NC, where I feel more area could be considered metro rather than rural. It appears I have also misclassified Grenada, MS.

I thought the UCC list would be quite helpful here. It's based on specific criteria measuring the urban vs rural population in the counties, and the urban fraction is listed for each county in the thread.

I'll definitely take a look at it later and do some comparisons. Hopefully a revision doesn't require too much additional tweaking!

[Adam Griffin]

Updated: I've revised all of the maps and spreadsheet data with the boundaries reflecting jimrtex's UCCs. Some small changes in the outcome:

• Obama won SC-Rural in 2012
• Trump won NH-Metro in 2016
• Trump won AZ-Metro in 2016

Margins of course changed in almost every state.

I'm having refresh issues with the animated GIFs on the forum showing the updated versions. If they're not updating for you (i.e. showing EC margins of 305-233 & 322-216), try opening them in a new tab.

[Gustaf]

Very interesting. I assume that the reason Democrats do better in rural areas in South Carolina and Mississippi is due to rural black voters. I wasn't aware this was so particular to those Southern states, compared to say Alabama?

[Gustaf]

Also, looking at your data I think you've flipped your 2012 and 2016 Diff columns!

[Adam Griffin]

Very interesting. I assume that the reason Democrats do better in rural areas in South Carolina and Mississippi is due to rural black voters. I wasn't aware this was so particular to those Southern states, compared to say Alabama?

Despite most Southern states having substantial swathes of rural black territory, none really come close to MS & SC in terms of sheer scope. Places like AL can be misleading on a map; virtually nobody lives in the (Democratic portion of the) Black Belt as a share of its state's population outside of Montgomery. The vast majority (around three-quarters) of MS residents live in rural areas in contrast. SC definitely is an anomaly, as about three-quarters of the population is in those metro areas. I'd also point out that in a good year and depending on the statewide election, Democrats can win a majority of square mileage in both MS & SC.

The Upper South (places such as TN) historically didn't have the kind of prime agricultural land in quantities sufficient enough to keep freed slaves and sharecroppers concentrated in numbers large enough to sustain communities in rural areas, which is why most black descendants in these states are heavily clustered in cities.

Even in most states where this wasn't the case (NC, GA, LA), sufficient urbanization and recent out-of-state black growth has led to both a purging of black residents from rural areas and/or massive growth concentrated in cities that makes said areas more Democratic than their rural counterparts.

MS & SC specifically don't fall into either of these categories. I'd argue SC does in the latter case to some degree, but rather than having one or more major urbanized areas, the feel of the state's clusters is definitely suburban - even by Southern standards of what constitutes "urban".

At any rate, these 2 states have rural areas that are blacker than their metro areas as a result, which makes them an anomaly by modern Southern standards - but historically in line with what once was. Of course, the figures tell the simple story of why these rural areas are more Democratic than their metro counterparts:

VAP:
SC-Rural: 59% White, 36% Black
SC-Metro: 69% White, 23% Black

MS-Rural: 60% White, 36% Black
MS-Metro: 62% White, 32% Black

AL-Rural: 74% White, 20% Black
AL-Metro: 67% White, 27% Black

Also, looking at your data I think you've flipped your 2012 and 2016 Diff columns!

Thanks for catching that: I've updated the sheet.

[Adam Griffin]

Thought I'd take a look at the polarization figures between metro and rural areas by Census region. Here are the results:

[StateBoiler]

Obama-Trump Districts:
ME-Rural
WI-Rural
MO-Metro

Romney Clinton Districts:
TX-Metro
AZ-Metro


And just in case anybody was wondering: the reason why the average EC margin is smaller than both the 2012 and 2016 EC margins is because Obama barely won MO-Metro in 2012, while Clinton barely won AZ-Metro & TX-Metro in 2016. In terms of average, all 3 of these jurisdictions ended up being won by the GOP by much more than the Dem won it by in the other election, thereby awarding their EVs to the GOP on the averaged map.

I am glad I took my biased lens off.. Missouri is going to be as Republican as Utah eventually!

Fun fact: MO-Rural had the largest margin in favor of Trump of any other state (Trump +54), with TN & KY (Trump +53), NE & OK (Trump +52) & WV (Trump +51) nipping at its toes!

haha yes! And last month in HD 144 there was literally a perfect Democratic candidate and a flawed Republican candidate in historically Democratic territory but the Democrat still lost by 5 points even if we managed to pick up an exurban seat.

If Missouri Democrats could get Platte and Clay County as solid democratic locally and Jefferson and St Charles as swing counties locally they may have a chance to compete state wide. But other wise its dead. Even if there maybe a last hurrah for constitutional officers in 2018/2020.

My feeling is that even if the state is elastic the suburbs are just too white and the rural areas are too southern cultured and too white to even be on the fringe of competitiveness.  It is not like downstate Illinois, rural Kentucky, rural West Virginia and all these other places where Trump did well yet local Democrats still held their own despite coattails. In Missouri it was a blood bath all the way down the ballot.

NE Indiana the same way. Looking at the 3rd congressional district:

Allen (Fort Wayne) - 50.5% of votes in the district:

Trump 83,867 57.4%
Clinton 55,379 37.9%
Johnson 6,801 4.7%

Other Counties - 49.5% of votes in the district:

Trump 107,723 75.4%
Clinton 25,869 18.1%
Johnson 9,328 6.5%

So when the two vote counts are roughly similar, Trump got 20% more votes in other counties, while Clinton got less than half. Her worst county she got 15.5% of the vote in (Johnson at 7.0% there), and all but 2 of the 11 she was under 20%.

[President Punxsutawney Phil]

Only 7 states in 2016 had a urban-rural differential of 50 or higher. TX was one of them.

[Adam Griffin]

Bumping this in light of recent discussions here.

[jamestroll]

Obama-Trump Districts:
ME-Rural
WI-Rural
MO-Metro

Romney Clinton Districts:
TX-Metro
AZ-Metro


And just in case anybody was wondering: the reason why the average EC margin is smaller than both the 2012 and 2016 EC margins is because Obama barely won MO-Metro in 2012, while Clinton barely won AZ-Metro & TX-Metro in 2016. In terms of average, all 3 of these jurisdictions ended up being won by the GOP by much more than the Dem won it by in the other election, thereby awarding their EVs to the GOP on the averaged map.

I am glad I took my biased lens off.. Missouri is going to be as Republican as Utah eventually!

Fun fact: MO-Rural had the largest margin in favor of Trump of any other state (Trump +54), with TN & KY (Trump +53), NE & OK (Trump +52) & WV (Trump +51) nipping at its toes!

We are literally a cycle away from Missouri being more Republican than Utah. I could see it happening in 2024!