Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

The House – new, with less democracy!

November 9th, 2012, 2:00pm by Sam Wang

Today, Speaker John Boehner stated that his party’s leverage comes from the fact that it retained control of the House. Yet they lost the popular vote. How can this be?

Before the election, I predicted that even if more people voted for Democratic House candidates, Republicans could still retain control. The reason I gave was redistricting since 2010, which has tilted the playing field significantly. The prediction was correct – though if anything, I underestimated the effect.

I estimated that Democrats would have to win the national popular vote by 2.5% in order to have a 50-50 chance of gaining control. I also predicted that the House popular vote margin would be D+0.0%, for Democratic gains of 2-22 seats . As of now, counting the leader in each undecided race, the new House will be 235 R, 200 D, a gain of only 7 seats. ThinkProgress reports a popular-vote tally of 50.3% D to 49.7%, a margin of D+0.6%. Both results are within range of my prediction.

However, this is quite notable. The popular vote was a swing of more than 6% from the 2010 election, which was 53.5% R, 46.5% D. Yet the composition of the House hardly changed – and the party that got more votes is not in control. This discrepancy between popular votes and seat counts is the largest since 1950.

Did I underestimate the tilt of the playing field? Based on how far the red data point is from the black prediction line, the “structural unfairness” may be higher – as much as 5% of the popular vote. That is incredible. Clearly nonpartisan redistricting reform would be in our democracy’s best interests.

Incidentally, some readers have suggested to me a reform in the Electoral College so that each Congressional district votes for its elector directly. As you can see, such a rule change would allow redistricting to influence the fairness of the Electoral College. Winner-take-all state races occasionally cause a problem, but which party gains is somewhat variable. It seems that there are worse things than the status quo.

Tags: 2012 Election · House · Redistricting

136 Comments so far ↓

  • Olav Grinde

    Good reason to question Arizona election results

    Did Flake really beat Carmona in Arizona?
    Did Romney really win Arizona?

    Jeff Flake received 960,318 votes or 49.8%
    Richard Carmona received 881,966 votes or 45.8%

    Romney received 1,071,557 votes or 54.3%
    Obama received 869,355 votes or 44.0%

    So what’s the problem?
    Well, as of Thursday 631,274 votes remained uncounted in Arizona!

    That is an astonishing number! A high percentage of this is provisional votes – disproportionately cast by minority voters. Much of the number is by voters who no longer found themselves on the voting rolls.

    It seems that 1) Arizona needs to finish counting, and 2) Arizona has some answering to do.

  • ess dee

    The problem isn’t that there was redistricting, it’s that there are districts. So long as artificial geographic boundaries define voting blocs, the gerrymandering will continue.

    To change the status quo (ongoing gerrymandering), we must dispense with geography and come up with a way of allowing citizens to self-identify which people they want to represent their varied interests. This might means giving each person multiple votes, or it might mean allowing people to vote for a party rather than a person. Or it might mean something else.

    But what I’m sure of is the speciousness of telling people that they are akin to all the other people in that geographic region by virtue of where they choose to live.

  • Mr. X

    Just tickled to see Dems whine about gerrymandering. They held the House for 40 straight years, even through two consecutive Reagan landslides, due to gerrymandering. I didn’t see them complain about gerrymandering back then.

    • Sam Wang

      I think your reading is inaccurate. Look at the cloud of data: its intercept is not far from the origin. Gerrymandering has become extreme in ways that have not occurred before. Before this election, only once since WWI has the minority vote-getting party ended up with a majority of seats – in 1996And that margin was only a few tenths of a percentage point.

  • Andrew


    Again, Gerrymandering appears to have become extreme because over the past 30 years, Democrats have gained urban and suburban voters and shed rural and small town voters to the Republicans. This has made Democratic urban islands nestled in a Republican sea of land.

    When one party’s voters are densely packed in a small area, they will typically end up in a handful of districts they win overwhelmingly, while the other party will win more numerous districts merely by a comfortable margin. The only way around this is the politically unpallatable solution of dividing up the Democratic cities into numerous districts. This tends to violate the Voting Rights Act by diluting minority voting strength and annoys most local elected officials, who want to interface with as few congressmen as possible in bringing their needs before Washington.

    In Pennsylvania, the Democrats failed to win close seats like CD 8 (Bucks County) and CD 12 (Cambria County) even though they lean Democratic and were won by Obama because 225,000 Democratic voters failed to cast a ballot for Congressional races.

    Philadelphia, for example, has slightly over two districts worth of voters, and so it is cut into three districts, 2 primarily in the city, and one mostly suburban. The whole Philadelphia region has 5.5 seats worth of voters, and is cut into 6 districts. Even if you sliced and diced the whole region to drop 200,000 to 450,000 residents of Philly into 4 seats and packed exurban Republicans into 2 of them, you would only change the present 3D, 1 Tossup, 2R map into a 4D, 2R map – virtually no difference. In other words, the seats won have little to do with any “Gerrymandering”, however squiggly the lines are.

    The situation cuts both ways too. In small heavily Democratic states where there are little or no geographic areas that are majority Republican, like Connecticut, Maine, New Hampshire, Rhode Island, Massachusetts and Maryland, Republicans are 40% of the electorate but hold just 1 out of 28 congressional seats instead of the 11 seats proportionality would suggest. These are simply the results of the distribution of voters across the landscape, not extreme partisan Gerrymandering.

  • Andrew


    Also, using your prediction line, if you move it up to intercept Democratic controlled congresses above the line, you will find that the line crosses zero at voting margins up to 5% favoring the Democrats. This is from past line drawing, especially in the south, that divided up black Democratic voters in order to produce untouchable Democratic margins in mostly majority white southern districts. The 1991 iterration of the Voting Rights Act mandated a change to this practice, and has lead to 8 of the last 11 congresses being Republican, where in 60 years before the Act, Republicans won a majority of seats in just 2 of 30 elections.

  • William Grisaitis

    Hey Sam!

    Quick question: what’s the data source for the pretty graph? Congressional Quarterly? If it’s a publicly available API or cvs or what not, I’d be terribly grateful to know.

    I’m a big fan of your work, btw. I actually found out about you through Nate Silver, namely his book.



  • Gary

    Though it is nice to have a local congressperson that can be called with your problem of the day, our gerrymandered system has forever been unfair. I propose that the house be elected on an at large basis. Of course this has been proposed numerous times and there are details to be worked out, but nothing insurmountable. If the will is there, the details can be worked out.

  • BNW

    Using Dave Wasserman’s tabulations for the House national Popular vote, Democrats’ lead has increased to D+.86% (from D+.6% when this article was written). The evidence for “structural unfairness” in House (re)districting is stronger than at the time of the original post.

Leave a Reply to William Grisaitis (Cancel)