The Myths of Red and Blue States

Readers of this blog might be interested in a new book on electoral politics set to arrive in bookstores at the end of summer. Red State, Blue State, Rich State, Poor State by Andrew Gelman — my colleague in Columbia University’s political science department — explodes some well-trod myths about American voting behavior.

Consider, for example, what he calls the “red-blue paradox”: rich states vote Democratic; poor states vote Republican. Two explanations that readily come to mind are shown to be invalid by Gelman: namely that Democrats are picking up a greater share of rich votes (not true); or that rich states are growing more socially liberal (also not true). To provide a better account, Gelman works his way, state by state, to help us better understand the relationship of class, culture, and voting. The book is a terrific read and offers much insight into the changing electoral landscape.

Taking class into account: at the lower end of the income spectrum Gelman doesn’t really find significant differences in voting between the states — in red and blue states, the poor vote similarly.

At the higher end of the income spectrum, however, the differences are more stark. In his words:

… income strongly predicts Republican voting in red America but not in blue America, where rich people are conflicted in their economic and social views … the key question is, what happened in the past twenty years to explain the red-blue pattern among upper-income voters?

The rich who live in red states have grown much more conservative on social issues (than their counterparts in blue states) and this trend has produced great gains for the Republican Party.

Perhaps this suggests that Obama’s political strategy should be targeted toward rich voters in blue states. That is, red staters are going to vote as they always have, so why spend the resources to change their minds. For his part, McCain ought to hold onto his blue state voters, since he (also) can’t do much to alter voting patterns in red states.

Elsewhere, Gelman writes that

Our results stand in contradiction to the commonly held idea that social issues detract lower income voters from their natural economic concerns.

Church attendance predicts Republican voting among the rich, not necessarily among the poor. And even more directly, he states:

It does not appear to be the case that rich people vote based on their economic interests, with lower income voters being more likely to be swayed by emotional appeals.

I wonder what Thomas Frank might say in response.


this book STILL doesn't explain what's the matter with Kansas


I presume it is safe to say that the rap music industry was not part of this survey. There I think you'd find it was ten times more, not ten percent less.


"The Republicon Guard must be p.o.’d that they were the red ones."

Actually, I like being on the "red" side. Red is a passionate color, a bold color, the color of the blood patriots spilled to keep our nation free.

Snot Rag Dave

"Church attendance predicts Republican voting among the rich, not necessarily among the poor."

Very subjective. In rural, predominately white communities... you can bet the church-attending poor vote Republican. Consider that so many Protestant churches distribute 'voting guides' prior to every general election... and those are often produced by right-leaning organizations.

There is also a tremendous influence by James Dobson's Focus on the Family and other pseudo-religious PACs.

I can't speak for the Catholic church... no experience. Perhaps another poster can enlighten...?

Johnny E

It was revealed recently that Tim Russert coined the whole red-blue thang when he drew a map on his whiteboard. The Republicon Guard must be p.o.'d that they were the red ones.

But most states have a really small majority either way so it probably doesn't matter much. The primaries showed that it was more a rural vs. urban thing.

I read a book a long time ago about the Senate being way skewed towards rural issues. I wonder if that's still true. You'd think urban areas would have more votes.


What is a red State? Is it Montana which has a Democratic Governor and 2 Democratic Senators but voted Republican for President the last few times? Or is it Minnesota with a Republican Senator and Governor but which has voted Democratic for President in recent elections?

Most would say that Montana is Red and Minnesota Blue. Why??


"Taking class into account: at the lower end of the income spectrum Gelman doesn’t really find significant differences in voting between the states — in red and blue states, the poor vote similarly."

Okay ... and how do they vote? Democratic or Republican? You never say.

Mike D.

What is this article doing here? One member of a university department plugging a book by another member? This is nothing but a commercial to help Columbia's Political Science Department feather its nest and promote a particular personality to the department's collective research interests. Academics being what it is, I would LOVE to be privy to the in-department politics behind this article. I'll bet money Gelman and Venkatesh are allies in some departmental skirmish over the department's vision for itself.

Susan H.

What I find most striking here is the notion that the American electorate consists entirely of "rich" and "poor" voters...almost no one outside of the most desperately needy or the most rarified elites would likely accept being labeled "poor" or "rich." Though there are no doubt far more poor than rich (which I would tend to think of as the upper 1% in terms of income), the vast, vast majority would consider themselves as "working" or "middle" class. Is it actually the middle class that votes based on religion in Red States? That would make more sense than "the rich," since there aren't enough of them to swing elections with their votes. Another characteristic of Red States is the racial aspect--the tendency for blacks to vote Democratic and whites to vote Republican (which is a recent historical trend traced to Nixon's "Southern Strategy"--the "red" states were solidly Democratic until the Democrats became the party of civil rights). Black and white seem to be far more relevant to voting patterns than "red" and "blue."



Convenient balderdash.

red-state-, blue-state- corned beef hash.

Sources say ....

The study finds ....

Rupert's editors mined it and refined it

into a little dissimulating lump of red-blue clay.


DJH: "Example: I live in a “blue” state, Connecticut, but in a portion of the state (the northwest) which is VERY “red.” How does the red/blue model account for that? Answer: It doesn’t. In fact, it DENIES the possibility of it!".

Nice try, but it does. If your state contributes to the electoral college as Democratic (Blue) votes, then you are a blue state. If there are enough true-blue voters, than it does not matter what the NW portion does - they do not affect the outcome. So, political plotters will skip your state, or the Dems will focus on all but the NW.

I am not saying this is a good thing, I am saying this Red/Blue thing is based around Presidential campaigns and the electoral college.

The Red/Blue divide is not Statewide for Representatives of course, only at voting districts (which are often screwy shapes based on gerrymandering). But, the same concept may apply.

The point is that there are enough voters who vote on party lines at the level of Fed gov, that this concept sadly holds true in many places. The outcome may be influenced by current situation (economy, war, fear, etc) or even the candidate (rarely), but it has to be enough of them to change the outcome, else the effort/money is wasted. The reasons tend to be largely due to the complexity of positions at the Federal level makes it hard for many to try to make a balanced decision. The old canards of "Democrats care about working class and people of color" and "Republicans care about family values and not taxing you" are enough to sway many voters to just find the "D" or "R" on the ballot and tick it.

At the local and State level, they tend to focus on the candidates and the local issues that matter to them, so the colors mean nothing or a lot less. That is, a red area may be socially conservative and will vote locally R or D based on who passes their social "litmus test". Another area may be economically conservative and will vote R or D based on taxation or trade issues. But, these issues do not play as easily at the Fed level (US Senators and Representatives, and Presidents) and so party line voting is far more common.


Pierce Randall

To me, this says that campaigning is going to have a tough time altering voting patterns against demographic change. The South, as heatedly discussed, is very Red state right now, but one day soon, enough people are going to live in Atlanta, Miami, Charlotte, Greensboro, and suburbs of Washington, DC to outweigh "upstate" voters in Georgia, Florida, North Carolina, and Virginia. That's mostly why you see Virginia shifting blue right now. Once the tipping point is reached in Sun Belt Red states, because rich city-dwellers are way more liberal than rich people outside of major urban areas, you're going to see these states shift and not look back.

Which is a good thing, in my opinion, because cities clearly get shafted by state politics almost everywhere in the country. In fact, most states seem to be jerrymandered "districts" that contain just enough agricultural/small town/rural interest to outweigh the technologica/industrial/cultural urban/surburan interest.



I suggest a rainbow state model that spans the continuous spectrum of electoral views in this country.

Violet California, New York, Oregon, Washington, Delaware, Vermont, ...

Blue Illinois, Massachusetts, New Jersey, New Hampshire, Wisconsin ...

Green Ohio, Nevada, Michigan, Arkansas, Iowa, ...

Yellow Florida, New Mexico, Virginia, North Carolina, Indiana ...

Orange Tennessee, Kentucky, South Carolina ...

Red Texas, Alabama, Georgia, ...

Johnny E

I would love to see the Freakos come up with a plan to audit the election before it happens so we know we can trust the results this time. Where is caging most likely to happen? How can we prevent it? How big an effect will the Indiana voter-id Supreme Court decision have on the presidential election? Has anybody really evaluated Greg Palast's statistics on caging? What percentage of votes (ie. provisional, military, absentee) never get counted?


I believe Gelman (and his co-authors in a previous paper by the same name: are using the "red state" "blue state" terminology not to suggest that these states are all inherently and uniformly politically different from each other, but instead to suggest that income matters differently in states that historically vote Republican versus those that historically vote Democrat. Their findings only suggest that income matters more in historically Republican states than in historically Democratic states. It also happens that Republican states are generally more poor. They therefore conclude that in poor/red states, rich people vote Republican and poor people vote Democrat, while in rich/blue states, rich people vote more Democratic.

I was hoping that Mr. Venkatesh's article would help to elaborate WHY this is happening. Educational differences between rich and poor states? Religious differences between them?



PaulK (#4), the red-blue state model is inherently flawed, and the reason for it has NOTHING to do with the identity or affiliation of the White House occupant.

Example: I live in a "blue" state, Connecticut, but in a portion of the state (the northwest) which is VERY "red." How does the red/blue model account for that? Answer: It doesn't. In fact, it DENIES the possibility of it!

That a state with millions of people is not ideologically homogeneous, should not surprise anyone. Even a tiny state like mine has more than enough voters to make homogeneity impossibile. Yet the red/blue model claims the opposite. It's irrational ... but pundits love it nonetheless.

That George Bush is the president has NOTHING WHATEVER to do with this reality. CT overall has been a majority-liberal state for decades; and northwestern CT as a portion of the state has been majority-conservative for at least as long, if not longer. Both of these facts were true before Bush was elected, and likely will remain true long after he is gone. Both would ALSO still have been true, had Bush never been elected at all.



"two newspapers in one"?!

I would like to refer you to today's NYTIMES opinion article by David Brooks that shows with numbers grouping of rich voters/donors:

republican: managers & old industries

democratic: professionals and high tech

PS I wonder whether democratic high-tech can be partially explained by age


The following quote interested me.

"It does not appear to be the case that rich people vote based on their economic interests, with lower income voters being more likely to be swayed by emotional appeals."

It is interesting because it might bring up an aspect that I haven't thought of and may or may not be brought up in the book. Does this mean that there is a correlation between emotion and politics? How about the emotion and being rich? Are people that can control emotion more successful in life and in business? Is this way Donald Trump always states “it’s not personal, it’s business”? Granted we are humans and emotions drive us, such as emotions for the environment (be green!), but do emotions get in the way of business and good business sense and does it affect a person’s wallet? If emotions do not get in the way of business deals do emotions get in the way in a person’s personal life, such as buying a house or other liabilities that they cannot afford that makes them poor?


Ben A

DJH (#1) pretty much hit the nail on the head with this one.

It's not I think we can't draw helpful generalizations from red-state/blue-state distinctions, it's just that I worry we've started to see such a rigid red/blue divide that it's started to cloud our view. As #1 said, the media don't really deal with the concept as a historical marker, but reather as some clearly establishable metaphysical property of a state.


Wow, thanks, MRB. I started out interested in an article about a more precise breakdown of voting results, and ended up reading an article by an unabashed hypocritical bigot.

"They--rural, red-state voters, the denizens of the exurbs--are not real Americans. They are rubes, fools, and hate-mongers."

Really? A person is a fool and a hate-monger based on the color of his state (sorry, county)? Incredible!

What people like this don't seem to realize is that the very fact that we have elections, and the candidates CAMPAIGN, is proof of one thing: people in this country are not painted red or blue.

This whole concept is upside down, in my opinion. A state BECOMES red, when the aggregate decisions of all it's voters add up to more votes for a Republican. It's an epiphenomenon. It makes no sense to investigate how people in a red state vote. That's like analyzing why people tend to build skyscrapers in places where there are lots of elevators.