Andrew Gelman's 2008 book Red State, Blue State, Rich State, Poor State: Why Americans Vote the Way They Do is an outstanding work of political science, rigorous yet accessible to the quantitatively inclined non-academic.

Along with uncovering much that is new regarding voting patterns, the book reaffirms some old truths that have tended to be glossed over in recent years by more impressionistic observers, such as David Brooks, with his sociological portraits of affluent coastal Democrats and economically humbler inland Republicans.

For example, Gelman finds that the party that emphasizes tax-cutting, the GOP, tends to appeal more to those who pay a lot of taxes. This is one of those stereotypes that is so obvious that it can get overlooked.

I would, however, warn that the catchy Dr. Seuss-inspired title--Red State, Blue State--can mislead readers into expecting the book to focus on explaining why some states on the electoral map are red (for Republicans) and others are blue (for Democrats). Ever since the 2000 election, commentators have been puzzling over why so much of the Electoral Vote map (and even more of the county-level map) is painted red for Republicans, even though the Democrats have done fairly well for themselves, winning the popular vote in two of the last three Presidential elections.

For example, here is Mark Newman's 2008 map where a county that 50-50 split between Obama and McCain is set to purple. Even in solidly blue 2008, the map is much more red than blue because Republicans do so well in low-density interior regions.

Instead, Dr. Gelman is most successful at explaining not the differences between states, but the differences within states. Within states, the rich are significantly more likely than the poor to vote Republican. This, he finds, is especially true in "poor" states with lower nominal incomes, which tend to vote Republican. (I put "poor" in quotes because the cost of middle class living, including homeownership rather than renting, varies so much between, say, blue California and red Texas due to differences in land prices that ranking them simply on nominal income can give a doubtful picture of their relative standards of living, as the current foreclosure crisis centered in California suggests.)

In other words, within red states we generally see the kind of old-fashioned relationship between income and voting that, say, Herbert Hoover's Treasury Secretary Andrew Mellon would instantly recognize: the richer you are, the more Republican you are.

Blue states, in partial contrast, now tend to feature a more diffuse relationship between income and voting (of the kind skewered in Christian Lander's Stuff White People Like). And yet, even in the bluest states, the rich tend to be more Republican, just not as much so as in the poorest states.

Those are very important insights, but they don't necessarily address what the title suggests the book is about: What colors states red or blue?

Consider this graph from the 2007 article "Rich State, Poor State, Red State, Blue State: What's the Matter with Connecticut?" in the Quarterly Journal of Political Science.

We see that the rich in impecunious Mississippi are more likely to vote Republican than the rich in affluent Connecticut. Yet, at most above-poverty levels of income, Mississippians are about 10 to 30 percentage points 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?

Popular answers to this question generally involve explaining that supporters of the other party are just plain deluded, stupid, or evil, stupid.

To answer the question of why inland areas tend to be redder than coastal areas, I've developed over the last decade a theory that ties a state's voting proclivities to its geography and its geography to its relative level of what I call "affordable family formation."

My approach is largely orthogonal to Dr. Gelman's, so we don't particularly contradict each other.

Dr. Gelman gives a summary of my Affordable Family Formation theory on pp. 170-171 of Red State, Blue State:

One link between economics, voting, and social attitudes has been noticed by journalist Steve Sailer, who hypothesizes that rich, coastal states now favor the Democrats because of increasing house prices, which reduces affordable family formation (marriage and childbearing), in turn limiting the electoral appeal of Republican candidates running on family values. Sailer attributes some of this home price difference to what he calls the Dirt Gap -- coastal and Great Lakes cities such as New York, Chicago, and San Francisco are bounded by water, which limits their potential for growth, as compared to inland cities such as Dallas or Atlanta: "The supply of suburban land available for development is larger in Red State cities, so the price is lower." The Republicans do better among married voters, who are more likely to end up in more affordable states that also happen to be more culturally conservative.

As a little thought experiment, imagine two sisters who are completely typical except that they are identical twins: with the same nature and nurture. They graduate from college together with degrees in business administration, but then they have to split up for the first time in their lives because one twin gets a job in downtown San Francisco while the other gets a job in downtown Dallas.

Ten years later, when they are 32, which twin is more likely to be a home-owning, married, stay-at-home mom?

Common sense and Census statistics suggest that the twin who got the job in Dallas is likely to take a more conservatively-inclined path through life. Middle class Americans today tend to get married when they are ready to buy a home and have children.

That houses are so much more affordable in the Dallas metropolitan area than in the San Francisco Bay region is one reason why non-Hispanic white women in Texas averaged 15.2 years of marriage between ages 18 and 44 in the 2000 Census, compared to 12.5 years for their counterparts in California.

And why is housing so much cheaper around Dallas than around San Francisco? There are many reasons, but a fundamental one is topographic: San Francisco is surrounded by saltwater and mountains, while Dallas is surrounded by flat dirt. There is simply a greater supply of land around an inland city than around a coastal city, so, ceteris paribus, the former's homes will be cheaper.

Dr. Gelman continues:

The geographic argument -- fewer families in coastal metropolitan areas because there is less room for affordable suburbs -- makes sense, even if it doesn't really explain why the people without children want to vote for Democrats and people with children want to vote for Republicans. It makes sense that more culturally conservative people are voting Republican, and these people are more likely to marry and have children at younger ages, but in that sense the key driving variable is the conservatism, not the family formation. It makes sense that the more culturally conservative people are voting Republican, and these people are more likely to marry and have children at younger ages, but in that sense the key driving variable is the conservatism, not the family formation.

In and endnote, Dr. Gelman supplies my answer: more family-oriented people will tend to move to more family-affordable places, but some others will be influenced by the local culture of wherever they happen to live.

Steve Sailer wrote about his affordable family formation hypothesis in an article, "Value Voters," in The American Conservative on 11 February 2008. He adds, "'family values' voters would thus tend to move to states where they could more easily afford a house with a yard in a satisfactory public school district, painting the Red States redder," and he speculates that "young people of middle-of-the-road tendencies might be more susceptible to starting down the path to conservatism in a state where family formation is quite affordable, such as Texas, than in an expensive state, such as California.

In a related article, "Why do gay men live in San Francisco," from the Journal of Urban Economics in 2002, Dan Black, Gary Gates, Seth Sanders, and Lowell Taylor argue that gay men disproportionately move to more attractive cities because they are less likely to have children and thus are more able to afford high housing prices."

Imagine it's now 20 years after college. The 42-year-old twin who got the job in San Francisco now has been married for five years, with a single three-year old daughter. She's working 25 hours per week to help pay the family's $4,000 per month mortgage on their two-bedroom house. Her twin sister in Dallas now has a 14-year-old son, an eleven-year-old daughter, and an eight-year-old son. Her husband's salary can pay their $1,600 per month mortgage on their four-bedroom house, so she stays at home and serves as her youngest son's den mother in the Cub Scouts.

One day, the twins are talking on the phone. One of them happens to mention a Democratic politician's campaign to ban the Boy Scouts of America from holding meetings on public school property because they won't accept gay scoutmasters. The Dallas twin with two sons in Scouting says: "I'm sorry, but it's only common sense that you don't let a gay man take your sons into the woods overnight."

The San Francisco twin with no sons is shocked by her sister's insensitivity, and replies, "But that's discriminating and discrimination can't be allowed."

The twins then get into a big argument that unnerves them both: How have they drifted so far apart in their fundamental values?

Well, life can do that to you. Different parts of the country are conducive to different ways of life, and our social and political attitudes tend to drift along with them.