The Divided States of America
By LEE DRUTMANSEPT. 22, 2016

Reading the nonstop coverage of what may well be a close presidential election, one might be forgiven for thinking that political competition is alive and well in America.

But look at the majority of states and congressional races, and a different picture emerges: In most places, meaningful two-party electoral competition is nonexistent. Rather than being one two-party nation, we are becoming two one-party nations.

Most large cities, college towns, the Northeast and the West Coast are deep-blue Democratic. Ruby-red Republican strongholds take up most of the South, the Great Plains, the Mountain States and the suburban and rural areas in between. Rather than compete directly against each other, both parties increasingly occupy their separate territories, with diminishing overlap and disappearing common accountability. They hear from very different constituents, with very different priorities. The minimal electoral incentives they do face all push toward nurturing, rather than bridging, those increasingly wide divisions.

Consider some numbers. The House, the supposed “people’s chamber,” is a sea of noncompetition. Out of 435 seats up for election this year, just 25 are considered tossups by The Cook Political Report. In 2014, 82 percent of House races were decided by at least 15 percentage points, including 17 percent that were not contested by one of the two major parties.

The Senate is only slightly better. A mere six seats out of 34 up for election are considered genuine tossups by Cook’s assessment (Florida, Indiana, Nevada, New Hampshire, North Carolina and Pennsylvania), while five are in the “lean” category.

The presidential candidates are also ignoring most of the country, instead focusing on the handful of swing states that always seem to take on outsize importance. In the 2012 presidential election, only four states were decided by five or fewer percentage points, and the median state-level margin of victory was a whopping 16.9 percent (in other words, not even close). Compare that with the 1976 presidential election, when 20 states were decided by five or fewer percentage points (and 31 were decided by eight percentage points or fewer), and the median state-level margin of victory was 5.9 percent.

While gerrymandering may explain some of the noncompetitiveness of House races, it can’t explain the Senate or the Electoral College. No amount of nonpartisan redistricting can overcome the fundamental disconnect between place-based, winner-take-all elections and polarized, geographically separated parties.

Competition is even rarer these days in state legislatures, where 43 percent of candidates did not face a major party opponent in 2014, and fewer than one in 20 races was decided by five percentage points or less. That made 2014 one of the most uncompetitive state-election years in decades.

These patterns are likely to continue: The current partisan geography is a natural political alignment. Around the world, urban areas tend to be left-leaning and cosmopolitan; rural and suburban areas tend to be conservative and populist.

This balance was thrown off for decades in the United States because the New Deal created an unlikely Democratic Party alliance between urban liberals and rural Southern conservatives (who were Democrats primarily because Republicans were the party of Lincoln and Reconstruction). This alignment could hold as long as civil rights was kept off the agenda. Yet, even after the turbulent 1960s, it took 50 years to cement the current arrangement.

One reason for the slow realignment in Congress was that in the 1970s and 1980s, many incumbent members of Congress mastered the art of constituency service, pork-barrel politics, personal self-promotion and enough fund-raising to scare off potential challengers. With weak party leadership in Congress, members were also free to “vote their districts,” even if that meant crossing party lines. This contributed to a blurring of partisan differences, which made individual candidates more important to voters than simple party labels. Partisan loyalties are also very, very sticky: Until 2010, Democrats held the majority in the Alabama State Legislature.

Perhaps the clearest measure of this slow transformation was the high rate of ticket-splitting that lasted for decades. From 1956 to 1996, on average 32 percent of House districts split their tickets, backing one party in the presidential race but the other for Congress. In 2012, that measure fell to 6 percent, the lowest since 1920 (3 percent). From 1968 to 1988, roughly half of states voted for different parties for the Senate and for president; in 2012, only six of 33 (18 percent) did.

As the parties became more homogeneous, rank-and-file members began to cede more authority to their leaders to enforce party discipline within Congress, especially in the House. Particularly after the watershed election of 1994, when many longtime conservative Democratic seats turned into relatively safe Republican seats, a new generation of conservative lawmakers and a newly assertive party leadership exerted a hard-right pull on the Republican Party. That election also bled the Democratic Party of many of its conservatives, shifting its caucus to the left. The election of 2010 was the culmination of the decades-long undoing of the New Deal coalition, sweeping away the few remaining Southern conservative Democrats.

Moreover, as more of the country became one-party territory, the opposing party in these places grasped the improbability of winning and so had little incentive to invest in mobilization and party building. This lack of investment further depleted a potential bench of future candidates and made future electoral competitions less and less likely.

These trends have been especially bad news for congressional Democrats, whose supporters are both more densely concentrated into urban areas (giving them fewer House seats) and less likely to vote in nonpresidential years (when most elections for governor are held, robbing the party of prominent state leaders). Since Republicans hold more relatively safe House seats, Democrats might benefit from occasional wave elections when the Republican brand has been significantly weakened (e.g., 2006 and 2008). But given the underlying dynamics, such elections are far more likely to be aberrations than long-lasting realignments.

An optimistic view of a future devoid of much electoral competition is that it saves members of Congress from having to constantly worry about re-election, which critics have argued pushes members toward short-term, parochial lawmaking. Perhaps all these safe seats can finally free up members to think beyond the next electoral cycle, and become genuine statesmen again.

Perhaps. But much would have to change in the current Congress to enable that kind of activity. At the very least, Congress would need to revitalize the committee and subcommittee system, and give individual members space to work across party lines and become long-term policy entrepreneurs.

The deeper problem is that a system without much effective two-party competition is a system that pushes entirely in the opposite direction. If a vast majority of seats are now safe for one party or the other, candidates don’t face any re-election incentive to reach out to the other party’s voters. Instead, they increasingly focus on the fear of a primary challenge. Especially on the right, this looming primary challenge (even if it is rare) means that members do not want to be seen as compromisers. This does not bode well for a functioning Congress.


By contrast, Congress was probably at its most fluid and productive during the periods of highest two-party competition, from the 1960s through the 1980s. This was partly because competition kept turnover steady enough that it brought in a relatively even flow of new members with new ideas. It also encouraged members to cut deals to bring home earmarks that would help them get elected.

Members don’t do these things anymore because they don’t have to. Whatever bipartisan bonhomie that once existed in Washington was a consequence of these underlying electoral conditions. Trying to re-establish that good will without fixing the underlying causes is like building a bridge across a river without foundations to ground the towers.

Certainly, there are some signs that we may have already hit the nadir of electoral non-competition. In presidential polling, for example, blue states are looking a little less blue this year than in past years, and red states are looking a little less red. Split-ticket voting will likely be up this year as well. If the Republican Party truly becomes the party of Donald J. Trump (and there is good reason to think it will), and Democrats continue to court moderate pro-business Republicans alienated by Mr. Trump while giving up on nostalgia-minded white working-class voters (also likely), this may make some states and congressional districts more competitive. Changing demographics, especially in places with rising immigrant populations, may also change the dynamics of competition. There are also some signs that divisions within the parties are coming to undermine longstanding party unity, creating potential for new crosscutting alliances in ways that are likely to reduce polarization.


But we have a long way to go. These nascent trends could use a boost. Perhaps we need to rethink our electoral model of winner-take-all elections, and particularly of single-member House districts — a model that reinforces, rather than cuts against, this growing geographic polarization, and one that makes it harder for parties to reflect their internal diversity.

The single-member, winner-take-all elections we use are a relative rarity among advanced democracies. They are not mandated by the Constitution, which lets states decide how to elect their representatives. In fact, many states originally used multimember districts. Returning to this approach would make it far easier to draw competitive districts that mix urban and rural areas. It would make it easier for different wings in both parties to send members to Congress, creating more diversity within the parties. It might also allow some smaller, regional parties to emerge, since multimember allow candidates to win with far less than majority support. These developments would increase the possibilities for deal-making in Congress. The FairVote proposal of multimember districts with ranked choice voting seems especially promising on this front.

But the first step in electoral reform is recognizing that this country has a problem. For decades, we had reasonably robust electoral competition, so there was little obvious reason to worry about our electoral system. But that era is over.

Lee Drutman is a senior fellow at New America and the author of “The Business of America Is Lobbying.”