Why Republicans Lost the Vote But Kept the House

Why Republicans Lost the Vote But Kept the House

By Sean Trende - May 16, 2013

One of the most striking aspects of the 2012 elections is that Republicans won their third-largest House majority since the late 1920s while losing the popular vote. Pundits have largely coalesced around a single explanation for this: GOP control of redistricting.

There’s no doubt that the party maximized its advantage by controlling redistricting in a majority of House districts, but that wasn’t really the culprit. The Democrats’ minority status has more to do with their “new coalition,” which might be good for winning presidential elections but is ill-suited for controlling the House.

Before wading into this, I want to emphasize that this whole discussion takes place within something of a false frame. The recent popular vote/House seat split is actually far less relevant than it seems at first blush, for three reasons. First, people don’t vote for “generic Republican” or “generic Democrat.” They vote for individuals. These candidates frequently fit the profile of the national parties, but not always.

Would the voters who cast ballots for Mike McIntyre, a North Carolina Democrat who often draws Focus on the Family’s endorsement, really have voted for “generic Democrat” if that were the choice? We’d have to call that possible but unlikely. So it’s hard to say that the American people tried to elect a House that would implement Barack Obama’s agenda. That agenda would still have faced considerable trouble if the Democrats had a four-seat advantage in the House, with McIntyre, Georgia’s John Barrow, Utah’s Jim Matheson, and Minnesota’s Colin Peterson holding the balance of power.

Second, right-of-center parties overall won 49.4 percent of the vote, just 0.1 percent of the vote less than liberal parties 49.5 (the remaining 1.1 percent split between various independent or 15th-party candidates). We can debate the meaning of votes cast for the “Overthrow All Incumbents” party, but the end result is that the public was actually pretty close to evenly split on whether they wanted a center-right or center-left coalition controlling Congress.

Third, the discrepancy between the popular vote won by a party and the seats that party wins in Congress is an endemic feature of our political system. The following chart shows the difference between the Democrats’ share of seats won in Congress and the share of the two-party vote they won over the past 70 years.

From 1942 -- the last time a party won the popular vote but failed to win the House (the Republicans won the popular vote by 5 percent that year) -- through 1992, the Democrats had a huge advantage in seats won vs. their popular vote share, averaging 5 percent. The discrepancy was less than 3 percent on just five occasions. The gap at times dwarfed the 4.5 percent difference between seats and popular vote share the Republicans currently enjoy.

Regardless, it is clear that the current district lines do work to Republicans’ advantage. The median district in the newly drawn Congress only went for Barack Obama by 2.5 percent in 2008 and actually went for Mitt Romney by 1.5 percent in 2012. This means that it is now impossible for Democrats to win the House without winning at least some districts carried by Mitt Romney (in fact, they have to win at least eight of them). In this sense, as Charlie Cook described it, the playing field is tilted against Democrats.

But do the Democrats’ troubles really stem from Republicans tilting the field? It’s difficult to isolate the effects of gerrymandering, but if we do, it becomes obvious that the problem isn’t the way the field has been tilted so much as that the Democrats are fielding a baseball team to play a football game.

One way to illustrate this is to test Republican redistricters against randomly generated maps. If the randomly generated maps have no bias to them, then we can more safely assume that redistricting efforts were behind the bias. But if these randomly generated districts have the same type of biases that we see in the Republican-drawn maps, then we have to look elsewhere for our explanation.

In fact, political scientists who have used computer algorithms to draw these types of random districts have generally found that they result in Republican advantages as well (the compact, randomly generated map of Florida, by the way, looks eerily like Florida’s current district lines).

But there are no proposals out there to draw congressional districts randomly. So instead, I wanted to test how Republican redistricters fared against maps theoretically drawn by independent commissions (the usual alternative to partisan lines). So, using Dave’s Redistricting App, I drew new lines for every state where presidential-level data was available and which had more than two districts.

I placed the following restrictions on myself: First, I drew the same number of minority-majority districts as were used in the post-2010 redistricting for each state (e.g., Georgia gets four African-American majority districts), although I endeavored to draw them as compactly as possible. Second, I constrained myself to “traditional redistricting criteria.” I didn’t consider where incumbents lived, and drew districts that were contiguous and compact.

To further minimize my discretion and eliminate any unconscious use of my knowledge of political geography, as well as to minimize county-splitting, I forced myself to draw the maximum number of districts in the most populous county in the state first, then the second-most populous, then the third-most, and so forth. This meant that I couldn’t draw a Pennsylvania district extending from Delaware County all the way up to the outskirts of Reading and Lancaster, as Republican redistricters did, because Delaware County is populous enough to accommodate its own congressional district. Within counties, I avoided splitting cities.

After all of this redrawing, Barack Obama won the median congressional district by four points against John McCain, an improvement of a little more than a point over the current lines. In other words, simulating independent commissions, we end up with roughly the same number of Republican-leaning and Democratic-leaning districts that we have today.

Nor does the analysis change much if we look at swing districts. Under the actual lines used in 2012, there are 41 districts that Obama won by between two and 12 points in 2008 (roughly five points less and five points more than his national margin), while under the redrawn lines there were 55 such districts.

Even expanding our definition of “swing seats” to include districts that Obama won by 17 points or lost by three doesn’t change things much. There are 101 such districts under the current lines, and 116 under the redrawn lines. If we assume Democrats would win these swing seats under the redrawn lines at roughly the same rate as they did under the current lines, they would pick up only six seats.

Why do Republicans maintain this advantage? We must remember that America’s congressional elections have always had a geographic basis. (Very early on, some states adopted “at large” statewide elections for their entire congressional delegations, but those became rare by the early 1800s.) There are drawbacks to this approach, but benefits as well: For one, it forces parties to form broad coalitions in order to win elections.

The Democrats’ problem -- something I first identified in late 2009 -- is that their coalition has become geographically narrow in the past decade, heavily concentrated among urban liberals and minorities who live in densely populated cities or are placed into minority-majority districts under interpretations of the Voting Rights Act that many minority groups pressed for in the 1980s and 1990s.

1 | 2 | Next Page››

Sean Trende is senior elections analyst for RealClearPolitics. He is a co-author of the 2014 Almanac of American Politics and author of The Lost Majority. He can be reached at Follow him on Twitter @SeanTrende.

Sean Trende

Author Archive

Follow Real Clear Politics

Latest On Twitter