Quantifying the great flip

By Razib Khan | August 30, 2012 1:13 am

The two maps above show the Democratic and Republican counties in blue and red respectively. Carter in the 1976 presidential election, and Obama in 2008. A few days ago it was brought to my attention that Matt Yglesias was curious about how Maine become a Democratic leaning state in the past generation. How is a deep question I’ll leave to political scientists, but how about the patterns of voting Democratic over elections by state for the past 100 years? That’s not too hard to find, there’s state-level election data online. So I just calculated the correlations between past elections and Democratic results, and Obama’s performance in 2008. If you’re a junkie of political science I assume you’ve seen something like this….


There is county level data out there, but it is hard to find the older stuff online. If you have some, especially the 1856 election, please contact me! If not, and you want to enter the data in by hand (you can find it in most libraries), I am willing to pay (I tried doing this once, but found it too boring).

Sorry if political posts bother you! But my Twitter feed and Facebook are starting to get inundated with dozens of posts every day on the perfidy of the Republican party. It’s hard to not think about politics a little….

MORE ABOUT: Politics
  • http://neuroskeptic.blogspot.com Neuroskeptic

    Very nice. What’s most interesting to me is that the big jump to a positive correlation happened with Kennedy, and then Johnson and Humphrey only added a little bit to that.

    Which goes against the narrative that it was Johnson and Civil Rights that created the modern political map…?

  • Danny

    the narrative that it was Johnson and Civil Rights that created the modern political map

    It was civil rights, sure, but not only Johnson. Already in 1948 there was an independent Dixiecrat in revolt against the incipient pro-civil rights tendencies in the Democratic party – that election already featured a positive correlation.

    This is an election with negative correlation. Looks pretty much like the inverse of the 2000 election.

  • Danny

    1976 was a bit of a throwback election – the Democrats won with a coalition of the South and Rust Belt – you can see it in the graph.

  • Dm

    Nice graph, and it appears to demonstrate a great deal of gradual continuity of demographic and cultural drift, rather than an abrupt “phase transition”?

  • http://dispatchesfromturtleisland.blogspot.com ohwilleke

    The political science term for the phenomena you are aligning is “Realignment” and is usually attributed to Goldwater and to Nixon’s “Southern Strategy”.

    Emerging out of the New Deal during which the Democrats became a dominant party nationally and not just in the South the near consensus fragmented into a de facto three party system – Northern Democrats, Southern Democrats and Republicans (mostly Northern). Starting with Goldwater this system gradually began to shift. In realignment, the Southern Democrats became Republicans, many of the modern Northeastern Republicans became Democrats, and the political system as a whole transitioned into one where there are two political parties that are relatively homogeneous nationally and are well defined by a simple one dimensional liberal-conservative dimension.

    By Obama’s election in 2008, realignment had pretty much run its course in federal elections (a number of Southern states have state and local level Democratic parties that retain their pre-realignment character to some extent). Only a handful of moderate Republicans and a few genuinely conservative Democrats are left in Congress.

    Notably, particularly in the context of this blog, the ideological leanings of voters at the county level has been much more stable, dating back to Reconstruction at that level and even back to the first couple of Congresses at the state level, than the political parties that implement those ideologies themselves.

    In 1876, the Republican party was the liberal party at the national level and the Democratic party was the conservative party at the national level. A century later, that parties had flipped ideological positions almost 180 degrees and the county level voting patterns had mostly flipped with those ideologies.

    This is bad news for someone out there trying to rock the vote as a campaigner (you are more likely to be successful in movement politics trying to shift the boundary of the line between liberal and conservative on a particular policy issue than to change the directional bias of the voters in a particular county), but it is fascinating for people who are interested in the long term stability of cultural pheneomena.

  • https://plus.google.com/109962494182694679780/posts Razib Khan


    1) both parties were less ideologically coherent between the civil war and the civil rights era. i.e., both had “progressives,” and there were “bourban democrats” and the wealthy republican establishment.

    2) some of the modern ideological fits are not useful. e.g., republicans were the part of high tariffs, and democrats free trade.. democrats more open immigration, and republicans more skeptical. but these positions were more reflection of cultural valences within the coalitions, not ideology as such. i think 50 years from now people will observe the same about us, that much that we presume is ideology is actually simply a reflection of alternative cultural consensuses.

  • Chad

    It should be noted regarding the Realignment:

    1) In Congress, it was the Republicans who were the primary supporters of the Civil Rights Act of 57, 60, and 64 while it was Democrats who primarily opposed it.

    2) If you look at Senate and Congressional Elections, there really wasn’t a Realignment (at least not in the South) until after Reagan.

    3) Nixon swept up the vote running against McGovern in 72, while in 68 much of the South went to Wallace, with Texas going to Democrat and California Republican.

    As a whole, if you look at all elections, Presidential, Senate, and Congressional, there was no clear realignment in the Nation until really after Reagan. Frankly, the facts don’t support the argument that it was Johnson and the Civil Rights act that redrew the map.

  • https://plus.google.com/109962494182694679780/posts Razib Khan

    #7, as you imply, “realignment” is conditioned on the granularity. national realignment (presidency) preceded the lowest office realignments by at least a generation. in parts of the deep south turnover of white nominal democrats for hyperlocal offices has only been occurring over the past 10 years or so. a lot of the southern white democrat congress people who change parties do so because of this disjunction; they are far more out of sync with their federal colleagues than they were in the state house of alabama.

  • Chad

    Razib, I agree that the lower offices were slower to switch over, but I really question the nature of the the Realignment, by which I mean I challenge the prevailing belief that it was the 64 election that really changed things.

    This data is no where as good, but if you simply look at how the states voted based on the maps in Wikipedia from 1960 up to 2008


    Its hard to really make the case that there was a realignment until 1992 and Clinton. It is only then do you start consistently see the pattern of Northeast and West Coast Democrat, Middle and South Republican.

    In 1980 and 84 Reagan pretty much sweeps the nation. In 88 the first Bush even got California. 76 with Carter resembles more the old traditional alignment than anything else. In 72 Nixon pretty much swept the nation. In 68 Nixon got places like California, but actually lost much of the deep South that went for Wallace, a Democrat who went rogue and ran as an Independent. In 1960 the South was still firmly Democrat. Its only briefly in 1964 that the South went Republican and places like California went Democrat. Otherwise all the years from 64 until Clinton defy the “Realignment”.

    Frankly, I think the current National Political map can be better explained by the Reagan and Clinton years and the battle between those two ideologies than the 64 Civil Rights story.

  • http://dispatchesfromturtleisland.blogspot.com ohwilleke

    “[T]he facts don’t support the argument that it was Johnson and the Civil Rights act that redrew the map.”

    There is a lot of inertia built into the political system. Incumbency advantages, stability in the ranks of status quo political elites, long staggered terms in the U.S. Senate. Johnson and the Civil Rights Acts were pivotal to starting the political movement that led to realignment, but it took a decade and a half for that movement to actually translate into electoral results.

  • Sam Schulman

    With all this talk about the South, look at New England: huge movement of population – yuppies in 80s into ME, NH, VT, CT – older Republicans retired and moved to FL. In general, 50s was first time middleclass people were rich enough to move en masse to Sunbelt for retirement, so the normal growth of conservative tendencies as an age cohort aged was excised. Also prewar+postwar move of industries from north to south – EG textiles – meant that Democratic party in north became less bluecollar, more professors, teachers, etc.

  • Dynotec

    The biggest thing people don’t realize is that something like 5-10% of the electorate of a state move every year. In the South Carolina Republican exit poll, more than half of voters were born out of state.

    Also, if you plot year to year correlations, one thing you’ll find is that things are shockingly stable on a state level.

    On a county level, it gets more tricky and complicated. One thing I’ve enjoyed is that the 1864 county map tracks the white vote today relatively closely.

  • Dan H

    A foreigner looking at the electoral map of 2008 would assume Republicans won in a landslide, rather than lost in a landslide.

  • Dan H

    The electoral map of 2008 is racial more than anything else. Blacks voted almost 100% for Obama and Democrats for racial reasons, even though a large proportion of blacks are religious conservatives. Blacks live overwhelmingly in and near large cities.

    Whites, although accused of racism much more than blacks, were far likelier to put substantive issues ahead of racial voting.

    Looking forward, it may be that blacks and increasingly hispanics are voting Democrat for purely racial reasons and regardless of ideology. Since these groups are increasing as a share of the population, the upshot is a possible future of Democratic dominance that is not issues or performance-based. This is not a sunny scenario; consider how the leaders of Detroit and some California cities have been able to govern with virtually no accountability, to disasterous effect.

    If Democrats can win elections regardless of performance due to racial voting patterns by minorities, an era of poor governance seems likely.

    Perhaps the future will see: (1) large numbers flood the Democratic party so they can at least participate, after racial voting by minorities make Democrats the only game in town, and (2) eventual splitting of the Democratic party into a new two-party system.

  • John Emerson

    There are a few areas which have remained Republican continuously while the country moved around them. New Hampshire is one. Wyoming is another. Maine’s two Republican Senators are relics of the old days. At the county level this would eb even more interesting because there were Republican pockets in the South which probably didn’t turn Democratic when the South shifted. There are also considerable areas of the South which remain diehard Democratic at the state and local levels while voting Republican for President.

  • https://plus.google.com/109962494182694679780/posts Razib Khan

    The electoral map of 2008 is racial more than anything else.

    have you done statisical analyses, or are we to rely on your omniscience? (it’s not too hard to do this analysis, i did a while back, and i recall that race is an important explainer, but the electorate is still 4/5 white i believe).

  • Anthony

    I reran the correlations based on how each state voted relative to the national vote, to minimize the effect of landslide elections, with data back to 1912 (n from 48 to 51). I also included third-party candidates. The strongest correlation to voting for Obama besides voting for a Democrat other than Carter from 1968 onwards was voting for John Anderson in 1980. Voting for Herbert Hoover correlates more strongly to voting for Obama than did voting for Carter.

    The other interesting thing is the variability of the vote, as measured by the standard deviation of the vote percentages for each state – since WW2, only 1948, 1964, and 1968 have had as much variation in state voting percentages relative to the overall national vote percentage as the 2008 election.

    Spreadsheet is here.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar