The “Incredible Shrinking Swing Seat” really does keep shrinking. In August 1997, The Cook Political Report introduced the Partisan Voter Index, an attempt to uniformly measure the competitiveness of all 435 congressional districts by comparing each district’s performance in the two most recent presidential elections with that of the nation as a whole. With the tireless help of Clark Bensen at Polidata, we have updated the PVI six times: after redistricting in 2002 and 2012, and after the presidential elections of 2000, 2004, 2008, and 2012.
By now, the trend lines are clear. In 1998, we found 164 swing seats—districts within 5 points of the national partisan average, with scores between R+5 and D+5 (a score of R+5 means the district’s vote for the Republican presidential nominees was 5 percentage points above the national average). The data 15 years ago showed just 148 solidly Republican districts and 123 solidly Democratic seats. Today, only 90 swing seats remain—a 45 percent decline—while the number of solidly Republican districts has risen to 186 and the count of solidly Democratic districts is up to 159.
In 1998, the median Democratic-held district had a PVI score of D+7, and the median Republican-held district had a PVI score of R+7—pretty partisan, but far from monolithic. Today, those median numbers are D+12 and R+10, and that 22-point gulf is the main structural driver of the political paralysis we lament today. Not coincidentally, the most Democratic and the most Republican House districts have never been further apart—Democratic Rep. Jose Serrano’s Bronx seat in New York City is D+43 on our scale, and Republican Rep. Mac Thornberry’s Texas Panhandle district is R+32—a 75-point chasm.
How did we get here as a country? Debates rage at political-science conferences: Are voters aligning in like-minded areas, or is blatant partisan gerrymandering to blame? Our newest index points mostly to the former, which has in turn amplified the power of the latter. In 2011 and 2012, redistricting diminished the number of swing seats from 103 to 99. But when we factored in the 2012 election results, the count fell more sharply, from 99 to 90.
In fact, setting aside redistricting, we found that 76 percent of Democratic-held House seats had grown even more Democratic in the past four years and 60 percent of GOP-held seats had grown even more Republican. Some districts swung dramatically, mostly along racial lines. Republican Rep. Hal Rogers’s 96 percent white district in eastern Kentucky jumped from R+16 to R+25, and Democratic Rep. Loretta Sanchez’s 82 percent nonwhite district in Orange County, Calif., moved from D+3 to D+9.
There’s no question this phenomenon has benefited Republicans. Observing Democrats trying to win a House majority today is akin to watching a soccer team play a comparably skilled opponent with the field slanted 25 degrees against them. Thanks to the concentrated nature of the Democratic vote, Republicans have always occupied at least two dozen more solidly partisan districts than Democrats. But the rise in ironclad districts forces Democrats to win a near-impossible share of what’s left just to even the score.
For Democrats, the additional bad news is that the range of what ought to be called a true swing seat might be shrinking. In 1998, Democrats held 15 of the 53 districts rated between R+2 and R+5, and Republicans held nine of 43 districts rated between D+2 and D+5. In other words, candidates on the wrong side of the “lean” of these “leaning” districts used to win a quarter of the time. But in 2013, Democrats hold just four of the 36 districts between R+2 and R+5, and Republicans hold zero of the 19 seats between D+2 and D+5—just a 7 percent crossover success rate.
By any measure, the House is now more sorted along partisan lines than ever. Using 2012 presidential results, only 17 Republicans represent districts carried by President Obama, and only nine Democrats represent districts carried by Mitt Romney. Using PVI scores, 18 Democrats hold districts that are more Republican than the national average, and only five Republicans hold seats that are more Democratic—a level of exposure that should sober Democrats who want to focus all of their party’s energies on offense in 2014.
Why aren’t voters splitting their tickets like they used to? Many theories, backed with compelling arguments, are floating around. But here’s one: As local print-news readership has declined precipitously in the Internet age, fewer voters are gaining exposure to individual candidates’ backgrounds and qualifications, and party labels have become an even-more-salient cue at the ballot box. To win on GOP terrain in 2014, Democrats need to localize a lot of races, but that has become much more difficult than it was just 10 years ago.
Our survey found one silver lining for Democrats in the new batch of PVI data: At the statewide level, 272 Electoral College votes—a majority—are now in Democratic-leaning states, to just 253 in Republican-leaning states, because Colorado shifted from Even to D+1 and Virginia moved from R+2 to Even. But at the congressional-district level, the tightening vise around swing seats is mostly squeezing Democrats.