As Americans examine the astounding dysfunction of their government, gerrymandering is usually cited as the prime culprit. This narrative offers a compelling villain: venal politicians who draw district boundaries for partisan advantage or to protect their own incumbency.
On the surface, it makes sense that manipulating district lines could be responsible for the increase in non-competitive, non-diverse congressional seats and the rise of ideologues who take radical positions without fear of voter retribution. But this ignores evidence that gerrymandering is only partly responsible for the current partisanship — and that eliminating it will not address the calamity we are witnessing.
No one disputes that congressional districts have become less competitive. During the last government shutdown in 1995, 79 of the 236 House Republicans represented districts that supported President Bill Clinton in his 1992 election. Today, only 17 of the 232 House Republicans represent districts that backed President Barack Obama — demonstrating more partisan consistency at the district level.
Cook's Political Report, a leading congressional handicapper, makes the point more directly. There were 164 competitive districts in 1998, according to Cook's Partisan Voter Index, but only 99 after 2012.
While this could be due to gerrymandering, a deeper look at the data reveals a different reality. County lines do not change every decade the way congressional districts do. In 1992, Clinton won 1,519 counties while in 2012 Obama won only 693 — less than one-third of all counties. His support was far more geographically concentrated than Clinton's.
This fact, more than gerrymandering, explains how Democratic congressional candidates could have over one million more votes in total than Republicans in 2012, but still could not win control of the House of Representatives. During the same period we have seen a decrease in the number of states considered competitive in presidential elections and fewer states sending split delegations to the Senate, even though state boundaries cannot be gerrymandered. Cook's data also shows that the number of competitive districts declined steadily during the last decade — not just after redistricting years.
Americans are more and more geographically segregated along ideological lines. When that segregation is refracted through the lens of party primaries and winner-take-all elections, the result is a Congress that is bitterly divided. Gerrymandering may work to make that division worse, but it is not the root cause.
Though public policy cannot easily address ideological segregation, all is not lost. Restoring the ideological middle ground to Congress is possible if we rethink how we organize our elections.
California and Washington State are now experimenting with "top two" primaries — where the top two finishers proceed to the general election. In theory, even when two candidates from the same party are selected in a heavily partisan district, the more moderate will attract voters from the other party and be elected.
The New York Times recently credited this reform for the relatively high functioning of the current California legislature. Unfortunately, empirical studies do not yet support this. Whatever effect "top two" is having comes at the cost of excluding political parties from the process of selecting candidates. In California, partisan affiliation is solely at the discretion of the candidate — relegating parties to the role of checkbook and a collection of personal endorsements.
At the same time, the system punishes multiple candidates with similar positioning, who run the risk of splitting the same vote and missing the cut for the general election. That's exactly the sorting problem that party primaries solve. A more promising alternative now used in several municipalities is ranked-choice voting (RCV). Under this system, voters rank multiple candidates for office. As the votes are counted, the candidate with the fewest votes is eliminated and his or her votes reassigned to the supporters' next highest choice. This process is repeated until a candidate has majority support among the remaining candidates.
At a minimum, RCV allows additional candidates to compete in the general election and have their arguments heard without the risk of being a "spoiler," giving voters the most choice at the election they are most likely to attend. At its best, it sends to office the candidate with the most support among all voters — not just the fraction that controls the party primaries.
The principal objection to RCV is the practical fear of voter confusion: 35 candidates filed for next month's Minneapolis mayoral race, where RCV will be used.
There is no such thing as a perfect voting system. What is clear, however, is that we have become locked in to a way of thinking about elections that no longer meets the best interests of our complex and divided country. It is possible, with ideas like these, to improve outcomes with simple changes in state law.
Gerrymandering is surely offensive to democratic ideals and should be eliminated. But if our goal is to restore competition to our elections and sanity to our politics, we need to think outside the lines.
(Jonathan Soros is a Reuters columnist. Opinions are his own.)