If there is one "politics" book you should read this year, it is Red State, Blue State, Rich State, Poor State: Why Americans Vote the Way They Do. Now, this sort of acclamation does need to be tempered by the fact that I myself don't really read "political" books very often. But despite the modest N, I'm rather confident that anyone who picks up Red State, Blue State, Rich State, Poor State will not be disappointed. To a great extent the collective of Andrew Gelman, David Park, Boris Shor, Joseph Bafumi and Jeronimo Cortina have produced a work which is a response in substance, if not style, to pundit productions by the likes of David Brooks and Thomas Frank. While the prose stylings of Brooks and Frank illustrate in an engaging manner what is "known," that is, fleshing out Conventional Wisdom with concrete specific exemplars, Red State, Blue State, Rich State, Poor State is more a series of charts girded by clarifying and interpretative text. Like good science Gelman et al. generate surprising and novel results, at times defying their prior expectations. This is a contrast to the modus operandi of most mainstream pundits, who select data which can vividly depict the plausible veracity of their hypotheses. While the core of Red State, Blue State, Rich State, Poor State is an analysis of the American political scene, their data sets suggest possible reasons that the pundits themselves come to the conclusions which they do. You not might agree with all the inferences made by Gelman et al. from their data, but there is value in clarity even if the contention is incorrect.
The basic argument of Red State, Blue State, Rich State, Poor State is that the assumption by many mainstream pundits that the socioeconomic elites tilt Left while the masses tilt Right is wrong. To a first approximation the rich tend to vote Republican, the poor tend to vote Democratic, the old stereotype holds. But there is also the reality that the wealthier states tend to vote Democratic and the poorer ones tend to vote Republican. In a book generously larded with charts I believe this is the most representative and informative one:
(cite: Quarterly Journal of Political Science Volume 2 Issue 4, 2007: "Rich State, Poor State, Red State, Blue State: What's the Matter with Connecticut?")
In short, as you can see, the rich in Connecticut, the wealthiest state per capita, are not much more Republican than the poor. In contrast, the rich in Mississippi, the poorest state, are much more Republican than the poor. Ohio, a middle income state, is somewhere in the middle. What Gelman et al. are showing here is that looking just at states removes critical information; class is a much better predictor of political orientation in poor states than it is in rich states. It isn't that rich states are blue because they are rich, it is that in rich states income doesn't matter much in relation to politics. You might wonder about the effect of race here; after all in Mississippi class and race are entangled and the political parties to a great extent are polarized along racial lines. Gelman et al. report that half of the effect is removed when controlling for race, but the general pattern of class based voting in poor states and not in rich states holds even for whites only.
By breaking down the data into county-level units Red State, Blue State, Rich State, Poor State also generates a plausible explanation for why mainstream pundit Conventional Wisdom would hold that the poor vote Republican more than the rich, even though on an aggregate national scale this is simply not so. In blue states the wealthier counties tend to be more blue (though within these counties the rich may still be more Republican; you can see this in Manhattan where the only precincts where Republicans attain parity with Democrats are in the wealthiest neighborhoods of the Upper East Side). Red states often exhibit an inverse pattern; the wealthiest counties are the most Republican! As it happens, the media elite is based in D.C. and New York, where the upscale regions tend to tilt heavily Democratic, while economically depressed rural counties in the hinterlands are more likely to tilt Republican. It stands to reason that David Brooks could easily confirm elite media perceptions that downscale locales tend to be more Republican and conservative than more affluent ones; the D.C. metropolitan area is a classic exemplar of the dynamics operative in blue America. In the end Gelman et al. are too gentle for my taste in criticizing the tendency of mainstream pundits, who no doubt fancy themselves cosmopolitans, generalizing from their own parochial existence, but on the merits they wipe the floor with them.
Though wealth gets mention in the title, I think it is important to highlight the fact that there is another independent variable which gets extensive treatment in Red State, Blue State, Rich State, Poor State: religion. "Religious State, Secular State" could easily have been tacked on if they wanted to create an even more ungainly title. It is likely no surprise that the more secular one is the more likely one is to vote for the Democratic party. This is a robust finding across states. In rich states, blue states, the rich are more secular than the poor. I doubt this will surprise many readers. But this might: in red states the rich are often more religious than the poor! In blue states the wealthy are more socially liberal than the poor, but somewhat more economically conservative. In other words, the wealthy in blue states tend to exhibit a libertarian lean. In red states the wealthy are at least, or more, socially conservative, and much more economically conservative. This is of course a natural explanation for why income matters much more in red states for partisan voting, but less so in blue states. Blue state elites have to balance their social values which are strongly Left with their economic interests. In red states elites have no such conundrum, their social and economic interests are naturally served by the Republican party, while the poor who may be socially conservative have powerful economic interests in the Democratic party.
To me one of the more interesting conclusions that Gelman et al. draw from these data is that this explains the power of economic conservatism despite the fact that public opinion polls tend to show that the public tends to tilt somewhat Left on these issues. While the Republican elites are unanimous in their support for economic conservatism, the Democratic elites are not in their hostility. "Pro-business" Democrats in states like Connecticut have to balance the fact that their constituents span the economic spectrum, from blue-collar union members to hedge fund managers with physics doctorates. Most readers are likely aware of the general outline of this phenomenon, but Red State, Blue State, Rich State, Poor State illustrates in stark relief the particular demographic reasons for this asymmetry between the two parties.
Though less than 200 pages of text, Red State, Blue State, Rich State, Poor State is a book of incredible substantive density. Despite its general academic lean it is written in an accessible manner with little technical jargon (e.g., you won't be confronted by constant references to beta coefficients, though those do exist in the end notes). Much of the material which Gelman et al. marshal can be found online, at the companion website which has a blog and archive of articles, or, at the personal weblogs of the authors. But Red State, Blue State, Rich State, Poor State takes all the appetizers and whips them into a very satisfying soufflé.
- Log in to post comments
Help me out here, razib. How does this Red State/Blue State (RS/BS)analysis compare with Steve Sailer's affordable family formation (AFF) hypothesis? The former shows that money matters less (ie, the rich are more likely to vote Democratic) in wealthy cosmopolitan areas, even though all else equal the wealthiest people still tend to vote Republican everywhere. Sailer's hypothesis focuses more on the poor, not in terms of their nominal family income, but rather in terms of the real affordability of land and housing issues.
I also notice that in red states like Mississippi where the upper fifth of the income distribution is more likely to vote Republican than in the coastal areas, the poor people are poorer relative both to their in-state elites and to the poor in other states. This raises the issue of whether voter participation might be lower among the poorer parts of the population in poor states (who would otherwise tend to vote Democratic), which might help turn them red.
Also, is religion more important in the lives of the bottom quintiles of the red state populations as compared to their economic counterparts in the blue areas? How does religiosity (say, church attendance) effect liklihood of voting among the poorer classes in these two sets of states? Is it different with Catholics v. non-mainline Protestant denominations for example? Are the church-going poor more or less likely to vote their pocketbooks than their unchurched brethern?
I guess I am thinking that Steve's affordable family formation hypothesis is maybe a more economical explanation (explains more with less) as well as being easier to intuit.
Sailer's hypothesis focuses more on the poor, not in terms of their nominal family income, but rather in terms of the real affordability of land and housing issues.
i think it would be more accurate to say it focuses on the middle class. really, the poor can't really afford family formation ;-) (or we can't afford their family formation)
Also, is religion more important in the lives of the bottom quintiles of the red state populations as compared to their economic counterparts in the blue areas?
yes, i think that "red" areas are more religious period, though in "blue" areas the evangelical and theological conservatives are also generally the poor (or lower middle class).
I guess I am thinking that Steve's affordable family formation hypothesis is maybe a more economical explanation (explains more with less) as well as being easier to intuit.
there are different ways to look at the same problem. sailer's model is mentioned relatively positively in the book. it seems that your recourse to possible different turnout rates in low SES groups in different states isn't very economical.
The thesis of this book is sometimes thought of as counterintuitive and startling, but actually it makes sense of things that the media more or less deliberately confuses. It's also consistent with the general principle that all movements (even "peasant rebellions") are led by elites of one sort or another, regardless of who the followers are.
Besides the religious / secular split within the relatively rich, I'd also mention the urban / non-urban split. But "urban" means "real cities" -- medium size cities (e.g. Wichita or Colorado Springs) can be extremely conservative. Beyond that there's the degree of attachment to the things you can get mostly in "real cities": things summed up as "culture" and "fashion". And still further on, there's a sort of continuum from the arts to the humanities to the social sciences to the sciences to the applied and tech areas. Likewise, as others have said, Democrats tend to rely on credentials and institutional affiliation for their living, whereas Republicans tend to be more directly exposed to the market.
Perhaps this book might help de-ideologicalize this question. I've been saying for a long time that Republican populism is fake, but Democratic elitism is real. That's not really a paradox. There's no law that says that the bottom half of the income distribution should have half the power. What the "law" says is pretty much the opposite: that the bottom half of the income distribution is often powerless.
Something I haven't studied but have run into a lot anecdotally (people I know or used to know, people I happen to meet) is prosperity theology. It's a pentecostal or charismatic kind of protestantism keyed to well-off people who might well be quite educated. A key phrase is "the Lord has blessed us". IE, success comes from The Lord. At best it's the old Christian idea that we are stewards of our wealth and are responsible to the lord to use it wisely and charitably. At worst it's the idea that well-off people have been rewarded by God whereas poor people are being punished by God. And there's a lot of magical thinking, since God's ways are unknowable and mysterious. Most secular people, at least the rational non-crazy ones, have secular explanations for their success or failure, whereas for prosperity Christians everything is part of God's mysterious plan.
I had a different thought when reading this:
This argument seems to treat wealth as the causative factor for ideology, yet ideology shapes what we do to such a large extent that it seems reasonable to propose that ideology is the causative factor for wealth.
Perhaps the blue states/regions are overall wealthier because their elites are more secular and use reason more in their decision making process.
John Emerson writes: "And there's a lot of magical thinking, since God's ways are unknowable and mysterious. Most secular people, at least the rational non-crazy ones, have secular explanations for their success or failure, whereas for prosperity Christians everything is part of God's mysterious plan."
Exactly. Perhaps the blue state secular elites are better at actually creating sustained and sustainable rationally-organized business ventures that result in higher employment and more wealth creation as a result of their more reason/reality-oriented ways of thinking.
I thought of one more thing to add.
I grew up in Ohio and now live in Massachusetts. Ohio is much more of a red state obviously, so I can offer some anecdotal observations on the effects of the dominant ideologies.
In Ohio, success is widely held to be the result of factors that are beyond one's control. I think this belief originates with religion, though it permeates the culture and is found among secular people as well. Success is held to be the effect of God's machinations or, if you're more secular, the result of birth, breeding, or chance.
Things like entrepreneurship, "pro-am" hobbies, inventing and tinkering, etc. are looked at as being silly by a substantial portion of the population. While nobody said it anywhere near this explicitly, the reaction that I received to some of my entrepreneurial thoughts and technological tinkering there was something along the lines of: "oh, you just haven't failed yet... when you fail you'll realize that you're powerless to affect your destiny and that people like you (not already rich) do not do things like this..."
In other words, mysticism and fatalism.
By contrast, when I talk about such things up here in heathen blue state Massachusetts I get excitement, offers to help, advice, etc. People here implicitly hold the idea that you can actually do things to affect your destiny, that learning is good, that tinkering or trying to DIY is not just a "silly hobby," etc.
In other words, reason and the belief that you can change things... the latter being the true timeless definition of liberalism.
This is, IMHO, the reason Ohio is poorer than Massachusetts despite the fact that there are many things it really has going for it.
Ideology shapes our actions, and our actions shape our world. Good ideas guide actions that yield good results, and poor and irrational ideas lead to poor and irrational actions or-- as in the case of the fatalism discussed above-- non-action and laziness.
My problem with books like this is that they try to treat ideology as a result rather than a cause. They do this to avoid broaching the topic of comparing ideology, since when you do that you must inevitably conclude that some beliefs/attitudes/philosophies/etc. are... well... better.
P.S. There are some beliefs and attitudes that I think are better in Ohio than in Massachusetts, but they are not sufficient to overcome the mysticism and fatalism of the culture. For one, folks up here seem to be more obsessed with credentials, background, etc. while Ohioan culture is in some ways more of a raw meritocracy. In other words there is, I believe, a strain of genuine elitism up here and that it's probably maladaptive. It results in people who are qualified on paper but not in practice getting promoted in the culture vs. people who really have the skills but maybe not as impressive of papers.
If you crossed blue state secularism and rationality with red state meritocracy and practicality you'd really have something.
Does the book address how race effects the white vote? For states where at least 10% of the population is black (excluding DC, for once a theoretical state has a black majority, it doesn't matter how whites vote), for example, there is a correlation of .42 (p=.07) between the state's Republican support and the relative size of its black population--that is, the blacker the state, the more likely it is to vote Republican. I suspect this is in part due to whites in heavily black states perceiving welfare programs as taking money from them and giving it to blacks, whereas in lilywhite states like Vermont, welfare programs are seen as merely taking money from overabundant whites and giving it to fellow whites who are momentarily down-and-out, but that's just a sloppy hunch.
I should back off a little from the magical thinking part. There's lot of secular magical thinking -- New Age, for example. A lot of the stuff people were saying about the stock market 2 or 3 years ago was magical.
The prosperity Christians I knew weren't necessarily irrational in the way they worked or the way they lived their lives. They just seemed to have this magical interpretation of the greater social world, such that their success was the result of God's blessing rather than, e.g., a real estate bubble.
My problem with books like this is that they try to treat ideology as a result rather than a cause.
i get a correlation of 0.00, when i check to see how political views relate to socioeconomic status and 0.26 in positive between conservatism and income. filter for race makes no difference.
. I suspect this is in part due to whites in heavily black states perceiving welfare programs as taking money from them and giving it to blacks, whereas in lilywhite states like Vermont, welfare programs are seen as merely taking money from overabundant whites and giving it to fellow whites who are momentarily down-and-out, but that's just a sloppy hunch.
the dynamic exists outside of southern red states with few blacks (e.g., great plains, rockies, etc.). a lot of this is south inflected, yes. the dynamic is most noticeable in the south, the class/republican voting pattern. but even in the non-south gelman et al. show the disjunction between red and blue states that is their general observation.
Right, but is that all they have to say about the racial angle? The authors report that half of the effect is removed when controlling for race--do they discuss the racial half, or just leave it at that and continue on with the class/geographic angles?
"Republican" vs. "Democrat" isn't ideology. Political parties are hodgepodges that try to put together 51% coalitions behind an array of local, state, and national candidates. They tend to tweak the coalitions until they approach the magic 51%, and as a result the ideological / policy content is very murky. (Example: under Clinton and the DLC the Democrats pulled away from labor and minorities and moved toward the middle class; after 1972 the Democrats locked the peace Democrats in the basement and kept them there.)
Even the "liberal" vs. "conservative" divide is particular to the US. American liberals are less socialist and more libertarian than the European left, and American conservatives are more authoritarian and more militarist that then European right.
So given how muddy the party lines are, it shouldn't be surprising if they don't correlate neatly with economic status.
In the Larry Bartels paper you linked to before the "white working class" was shown to be favorable toward conservative economic policies and liberal social policies. Stephen Rose of the EPI on WWTMK said much of the appeal of more conservative economic policy is "aspirational", and there aren't enough poor whites (or the working class is not poor enough) for Huey Long type politics to get much grip.
The Thomas Frank/David Brooks type spiel seem so embedded in our thought that even after Ilkka Kokkarinen linked to my post debunking it he kept repeating it himself.
do they discuss the racial half, or just leave it at that and continue on with the class/geographic angles?
yeah, they talk about how it might explain the extremity of the pattern in the south. as indicated by the title of the paper cited, their focus is disproportionately on how income does not predict voting patterns by party strongly in states like connecticut.
Political parties are hodgepodges that try to put together 51% coalitions behind an array of local, state, and national candidates.
right. but, at the elected official level they're much more politically polarized than they were. on the voter level they're considerably more partisan than they were. a lot of this is due to the decline of conservative southern democrats and liberal northeast republicans.
American liberals are less socialist and more libertarian than the European left, and American conservatives are more authoritarian and more militarist that then European right.
well, militarist mebee, but i don't know about authoritarian. the right in europe (and canada) is economically much more liberal, but there's a reason that germany has christian instruction in schools, much stronger restrictions on abortion than the USA, or that out-and-out racial nationalist parties can rack up 5-30% of the vote depending on the context. one might model it by diving the center-right from the radical right; the latter is far less the party of business and establishment social conservatives, ergo, the open appeals to xenophobia and class warfare. in the USA these two elements are in the same party, explaining the internecine conflicts over immigration and trade (which the econ conservatives always win).
According to WSJ the rich do support McCain, but the super-rich support Obama.
I've only read summaries of Gelman's work, even though he has been very kind toward my work, so I'm commenting kind of blind here.
I suspect we aren't looking at the same question, though. My Affordable Family Formation theory is about the predictability of which states will vote Democratic and which will vote Republican.
Gelman's big insight, in contrast, is about something in common across all states within each state -- the richer you are, the more likely to vote Republican. That's a useful thing to remember, but it doesn't seem to address what the title of the book suggests the book is about: what colors states Red or Blue?
For example, in the 2004 election, the slope of the line in the graph above was almost the same for Mississippi and Connecticut. The two lines were almost parallel -- but at each level of income, Mississippians were about 20% to 30% more likely to vote Republican than Nutmeg Staters. Now, _that_ is what strikes me as the real question: what's so different between the two states?
Leaving aside the race demographics for the moment, one big difference is the Affordability of Family Formation: housing costs far less in Mississippi especially near jobs. Lots of people send their kids to private schools in both states, but Stonewall Jackson Christian Academy in Tupelo is going to cost far less than Olde Grottlesex School in Litchfield, etc.
This is all very interesting, but what do Gelman et al. have to say about what I think is the most prominent source of variance in voting patterns today, which is sex? I believe that women overwhelmingly vote Democrat, whereas men are much more likely to vote Republican. Perhaps it hasn't always been this way, but do the authors address sex?
This is all very interesting, but what do Gelman et al. have to say about what I think is the most prominent source of variance in voting patterns today, which is sex?
race and religion exhibit much bigger variance between classes than sex, so this is not really totally accurate (though sex might account for the biggest component of variation, cuz men and women are 55 and 45 percent of the electorate, while blacks are only 10 percent). and i believe that the sex difference is the same across the trends they are interested in. as others have pointed out, the big reason that the sex gap exists is that single women vote overwhelmingly for democrats.
The marriage gap is usually bigger than the gender gap. The Democratic pollster Stanley Greenberg did a multiple regression analysis on the 2004 exit poll data and found that, all else being equal, a single person was 1.56 times more likely to vote for Kerry than a married person.
That's big.