The Big Sort (2 page)

Over the past thirty years, the United States has been sorting itself, sifting at the most microscopic levels of society, as people have packed children, CDs, and the family hound and moved. Between 4 and 5 percent of the population moves each year from one county to another—100 million Americans in the past decade. They are moving to take jobs, to be close to family, or to follow the sun. When they look for a place to live, they run through a checklist of amenities: Is there the right kind of church nearby? The right kind of coffee shop? How close is the neighborhood to the center of the city? What are the rents? Is the place safe? When people move, they also make choices about who their neighbors will be and who will share their new lives. Those are now political decisions, and they are having a profound effect on the nation's public life. It wasn't just my neighborhood that had tipped to become politically monogamous. In 1976, less than a quarter of Americans lived in places where the presidential election was a landslide. By 2004, nearly half of all voters lived in landslide counties.

In 2004, the press was buzzing about polarization, the inability of the leaders of the two political parties to find even a patch of common ground. All the measures of political ideology showed widening divisions between Democratic and Republican political leaders, and unbridled partisanship in national politics became a topic for Sunday news shows and newspaper columnists. Meanwhile, unnoticed, people had been reshaping the way they lived. Americans were forming tribes, not only in their neighborhoods but also in churches and volunteer groups. That's not the way people would describe what they were doing, but in every corner of society, people were creating new, more homogeneous relations. Churches were filled with people who looked alike and, more important, thought alike. So were clubs, civic organizations, and volunteer groups. Social psychologists had studied like-minded groups and could predict how people living and worshiping in homogeneous groups would react: as people heard their beliefs reflected and amplified, they would become more extreme in their thinking. What had happened over three decades wasn't a simple increase in political partisanship, but a more fundamental kind of self-perpetuating, self-reinforcing social division. The like-minded neighborhood supported the like-minded church, and both confirmed the image and beliefs of the tribe that lived and worshiped there. Americans were busy creating social resonators, and the hum that filled the air was the reverberated and amplified sound of their own voices and beliefs.

This was not an area of concern for most of those who wrote about politics. Migration wasn't thought to be much of a factor in politics. People moved, sure, and some states gained votes while others lost. But the effects were thought to be essentially a wash.
²
Frankly, I only stumbled upon this trend in American politics—and that was only after I stumbled upon Robert Cushing.

I had previously worked for a small paper in the coalfields of Eastern Kentucky, and my wife and I had owned a weekly newspaper in rural Texas. From my experience living in small towns, I had become interested in why some communities develop vibrant economies while others stagnate, and I had written about this question as a newspaper columnist in Kentucky and then a reporter in Austin. Cushing was a sociologist and statistician who had recently retired from the University of Texas. My parents were friends with a cousin of Bob's wife, Frances. Through that tenuous connection, we met for breakfast one morning.

I remember telling Bob I had some data about Austin's economy but didn't know quite what to do with it. "I do," Bob responded. That was typical Bob, a guy who had paid his way through graduate school by working summers fighting forest fires as a Smokejumper in Montana. He did know what to do with the pile of data I had collected, and we began collaborating on projects for the
Austin American-Statesman.
We would decide on a question we wanted to answer, and Bob would begin clicking, programming, and calculating. Often in the middle of the night, a new set of charts and Excel files would arrive in my e-mail in box, and I'd see that Bob had made another remarkable discovery.

Our interest initially was why a small group of cities, Austin among them, were growing so fast and so rich. In 2002, we began working with a band of researchers, including Richard Florida and Kevin Stolarick, then at Carnegie Mellon University; Gary Gates at the Urban Institute; Joe Cortright in Portland, Oregon; and Terry Nichols Clark at the University of Chicago. What we found was that these tech-rich and innovative cities were benefiting from a special kind of migration.

There have always been patterns to migration and development. Southern blacks moved to Chicago in the 1950s. White Appalachians took the "hillbilly highway" north to booming Cleveland and Detroit after World War II. These were migrations in response to economic hardship and opportunity. The movements we saw from 1970 to 2000 were different. The flows were selective, and they varied by personal characteristics, not broad demographic descriptions. People were sorting, and the movements themselves were changing economies. Young people left old industrial towns and rural America for urban life, moving to a small group of cities. People earning higher incomes went to Austin or San Francisco. Those with lower incomes moved out of those towns. Las Vegas attracted people with less education than those who went to Seattle. Portland slurped up people with college degrees. Blacks went to Atlanta and Washington, D.C.; whites went to Austin and Orlando.

Everyone in our research group had a particular specialty. Demographer Gary Gates discovered that gay couples were congregating in particular cities—specific Zip Codes, in fact. Rich Florida and Kevin Stolarick calculated the correlation between particular types of workers—what would become Florida's "creative class" of artists, engineers, software writers, and teachers—and economic growth.
³
Joe Cortright was interested in the migration of the young. Terry Clark discovered the amenities that attracted high-wage workers. Bob Cushing and I used patents to measure economic innovation in cities. We tracked people and their incomes from county to county using Internal Revenue Service (IRS) data. We used polling data from an advertising firm to find a relationship between culture and economic growth, and we tracked various demographic groups as they moved into high-tech cities such as Austin. Everyone in our group found movements that were dizzying and profound. We could see that distinct migratory streams were reshaping regional economies. Somebody in our little ad hoc research outfit said that the zigzagging of people and money looked like "the big sort," and the name stuck.

Initially, Cushing and I were so interested in the economic effects of the Big Sort that we paid no attention to whether the phenomenon meant anything to politics. That was an oversight. After all, the movements we were tracking were enormous. More people were moving from one county to another in a single year than new population was added nationally in four years. It was clear the Big Sort wasn't random. The sorting we saw in the last decades of the twentieth century created prosperity in some places and decimated economies in others. It just made sense that this internal migration had changed the nation's politics, too. Then I received one of those magnificent midnight e-mails from Cushing.

He had access to presidential voting results for each U.S. county since 1948, collected by the web-based election data impresario David Leip. The presidential election is the only national office common to all voters. We decided to use presidential election results—instead of either voter registration or state elections—as the common measurement among the nation's more than 3,100 counties to avoid the effects of different candidates or changing voting districts. To even out the comparisons over time, we excluded third-party candidates.
⁴
Demographers have several ways to measure segregation, and Cushing tested them all. The formulas tell, for example, whether the proportion of blacks and whites living in Brooklyn is similar to the proportion of the races in New York City as a whole. Substitute Republican and Democrat for black and white, and these formulas provide a measure of political segregation. We ran the numbers for all the counties in the United States, and a pattern emerged.
⁵
From 1948 to 1976, the vote jumped around, but in the close elections, Republicans and Democrats became more evenly mixed, especially in the 1976 contest (see
Table 1.1
). After 1976, the trend was for Republicans and Democrats to grow more geographically segregated.

The simplest way to describe this political big sort was to look across time at the proportion of voters who lived in landslide counties—counties where one party won by 20 percentage points or more. There were elections when the entire country seemed to side with one party or the other. In the electoral blowouts of 1964, 1972, and 1984, close to six out often voters lived in landslide counties. Landslide elections produced a lot of landslide communities.

Competitive elections provided a more accurate picture of where Republicans and Democrats were living. The 1976 election between President Gerald Ford and Jimmy Carter took place at the time after World War II when Americans were most likely to live, work, or worship with people who supported a different political party. Just over 26 percent of the nation's voters lived in landslide counties.

Then the country began segregating. In 1992, 37.7 percent of American voters lived in landslide counties. By 2000, that number had risen to 45.3 percent. There was a difference between the elections prior to 1976 and those that came after. In the polarizing, and close, 1968 election between Richard Nixon and Hubert Humphrey, 37.2 percent of voters lived in landslide counties. The last five presidential elections have all had a higher percentage of voters living in landslide counties than in 1968. Beginning in 1992, the percentage of people living in landslide counties began an upward, stairstep progression. And by 2004, in one of the closest presidential contests in history, 48.3 percent of voters lived in communities where the election wasn't close at all.

 

Table 1.1 Close Elections, Local Landslides

 

In competitive elections, more voters are living
in counties with lopsided presidential votes.

 

Note: This table shows the proportion of voters living in counties where the local presidential election was decided by 20 or more percentage points. Competitive elections are those decided by 10 percentage points or fewer. (Third-party votes aren't included.)

Source: Dave Leip's Atlas of U.S. Presidential Elections,
http://www.uselectionatlas.org.

 

To get a sense of the magnitude of this segregation by political party, we asked John Logan, then at the State University of New York at Albany, to measure the segregation of blacks and whites from 1980 to 2000 in all U.S. counties. In those twenty years, counties had become slightly less racially segregated. Using the same formula, however, from 1980 to 2000 the segregation of Republicans and Democrats increased by about 26 percent.
⁶
The trend didn't appear to be connected to the seesawing fortunes of the national parties. The increasing political segregation of American communities continued on through the Reagan landslides, the Democratic turnaround in 1986, the Clinton resurgence, the Republican takeover of Congress in 1994 and the Bush wins in 2000 and 2004. Moreover, this trend—one that was particularly strong beginning in the 1990s—had escaped the attention of those who study and write about politics.
⁷

The one political scientist who best understood the power of migration and the importance of community to politics was James Gimpel of the University of Maryland. Gimpel had written a book about the different migration paths followed by immigrants and U.S. citizens. Gimpel argued that migration was shaping communities and altering their politics. And in 2003, he wrote another book showing states to be "patchworks" of quite different and sometimes quite homogeneous communities.
⁸
Gimpel's work gave us confidence that migration could be shaping the complexion of a community's politics. What nobody had seen, however, was that most of America and most Americans were engaged in a thirty-year movement toward more homogeneous ways of living—that the polarization so apparent in the way political leaders talked was reflected in the way Americans lived.

Bob Cushing and I published these initial findings in a series of newspaper articles in the summer and fall of 2004. In the months after the 2004 election, we reconsidered what we had written for the paper. We knew that the pattern was clear. But was it meaningful? Did it matter that our communities were becoming more politically homogeneous? The country was polarized during the Civil War, but compared to those times, our own circumstances didn't appear dire. Furthermore, assuming that these demographic shifts
did
matter, why would they be happening now?